Tuning functional properties: From nanoscalebuilding blocks to hybrid nanomaterials

K GEORGE THOMAS

Photosciences and Photonics Group, National Institute for Interdisciplinary Science and Technology (CSIR)(formerly Regional Research Laboratory), Trivandrum 695 019, India.

andJawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560 064, India.

e-mail: kgt@vsnl.com

Materials in the nanometer dimension possess interesting functional properties originating mainlyfrom the quantum confinement effect, which is a beautiful experimental demonstration of the‘particle-in-a-box’ model. Confinement of charge carriers in nanoscale materials leads to dramaticmodifications in their density of states giving rise to size and shape dependent properties. Theability to tune the functional properties of materials without changing their chemical constituentshas opened up newer opportunities in disciplines as diverse as physics, chemistry, biology, medi-cine and engineering. The current focus of nanoscience is to design materials with novel propertiesand utilize them for the miniaturization of devices with better performance. Among the vari-ous nanoscale building blocks, metal and semiconductor nanoparticles and carbon nanotubes havegained much attention and a brief summary of their functional properties is discussed. Further-more, the functional properties of nanomaterials can be fine-tuned by a stepwise integration ofthese nanoscale building blocks into organic–inorganic and inorganic–inorganic hybrid systems.In the integrated hybrid nanomaterials, the functional properties of the individual nanoscale build-ing blocks can couple with each other to yield newer properties that are fundamentally differentfrom that of isolated components. The newer optical properties of nanomaterials provide excellentopportunities in the biomedical field for diagnosis, imaging and therapeutics. Hybrid nanomaterialswith diverse functionality find application in the design of new generation nanophotonic and opto-electronic materials, and these aspects are discussed.

1. Introduction

Controlled organization of materials with micro-meter scale precision has led to the miniatur-ization of devices, which has been the key tomajor technological innovations in the last decade.Design of materials in the nanometer scale is notmerely another step towards miniaturization; itis the exploration of a new size regime whereinmaterials exhibit properties which are differentfrom that of the bulk. In this regime, sizeand shape play a dominant role in tuning theirproperties rather than the constituting elements

and their chemical composition [1–10]. Exam-ples of nanostructured materials include metaland semiconductor nanoparticles and nanostruc-tured carbon based systems such as fullerenes,single and multi-walled carbon nanotubes. Thenoble metal nanoparticle based systems possess anattractive property in the nanoscale: strong opti-cal absorption arising from the localized surfaceplasmon resonance (yellow, red and blue color forAg, Au and Cu, respectively). The larger surfaceto volume ratio is yet another unique feature ofnanostructured materials, which make them a suit-able candidate in catalytic processes [1].

Keywords. Nanomaterials; metal nanoparticles; quantum dots; carbon nanotubes; hybrid systems; photochemistry; elec-tron transfer; energy transfer; plasmons; heterojunctions.

53

54 K GEORGE THOMAS

Nanoscale materials exhibit size and shapedependent physical, chemical, electronic and mag-netic properties, which are different from the bulkand their isolated atoms/molecules. Classical lawsof physics fail to explain the origin of the novelproperties of materials in this size regime [11].Electrons experience a confinement of motion inspace, when one or more dimensions of a crys-tal are in the nanoscale. This situation can bedescribed in terms of particles in a box and underthis condition, quantum mechanics is more suitableto explain the size and shape dependent properties(vide infra). Dimensional confinement of electronsin materials leads to dramatic modifications intheir density of states (DOS) giving rise to shapedependent properties. Based on dimensional con-finement of electrons, nanomaterials are generallyclassified as 2D, 1D or 0D [10]. The confinement ofelectrons is imposed only along one axis in the caseof 2D nanomaterials (e.g., quantum wells). In thecase of one-dimensional nanomaterials, electronsexperience confinement along two dimensions andare free to move only in one dimension; the bestexamples are semiconductor nanowires and carbonnanotubes. When the material experiences confine-ment of electrons along all the three dimensions, itis often termed as quantum dots (QDs). The dra-matic blue shift observed in the absorption andemission spectra of semiconductor quantum dots,with decrease in size, is one of the direct observa-tions of quantum size effect (vide infra).

The synthesis, characterization of various nano-materials and investigation of their properties arewell documented in recent reviews [1–10]. Newermethods of characterization based on electron andprobe microscopic methods have enabled betterunderstanding on the crystallographic propertiesof nanomaterials. Various spectroscopic methodshave been used for understanding the optical andelectronic properties of these materials. Amongvarious nanomaterials, gold nanoparticles andnanorods, semiconductor quantum dots and sin-gle walled carbon nanotubes are widely used ascomponents for the design of hybrid nanostruc-tures. This article provides a brief summary ofthe recent understanding on the functional pro-perties of nanomaterials, particularly on the opto-electronic properties of nanoscale building blockssuch as metal nanoparticles, semiconductor QDsand carbon nanotubes and hybrid nanomaterialsbased on these systems.

2. Nanomaterials and theirfunctional properties

2.1 Gold nanoparticles

Metal particles in its colloidal state have attractedmankind even centuries ago due to their medicinal

Figure 1. Schematic representation of the oscillation ofthe electron cloud in presence of an electromagnetic radia-tion (reproduced with permission from [8], copyright 2003American Chemical Society).

value and fascinating colors [11–14]. Colloidal goldhas a long therapeutic history, which is well rootedin Eastern traditions particularly in the Indian sub-continent [11–13]. The medicinal value of colloidalgold is well documented in the books of ancientIndian ayurveda like ‘Charaka Samhita’ and the‘Vedas.’ The fascinating colors of metal nanoparti-cles have been utilized for decorating glass windowsin many cathedrals in Europe (stained glass win-dows). Another interesting example is the famousLycurgus cup of 4th century, which is now dis-played in the British Museum [14]. This glass cupappears green when viewed in reflected light andtransmits red color, when illuminated from inside.Analysis of the glass reveals that it contains smallquantity of an alloy of gold and silver having adiameter of ∼70 nm in the molar ratio of (3 : 7).The surface plasmon resonance (vide infra) of thealloy is responsible for the special color display inthe Lycurgus cup.

The light absorption in metal nanoparti-cles originates from an interesting phenomenoncalled localized surface plasmon resonance [1–10].According to the Drude–Lorentz model, the atomsin metals exist in a plasma state, having a coreof positively charged nuclei surrounded by a poolof negatively charged electrons and hence namedas ‘plasma electrons’. In the presence of an elec-tromagnetic radiation, the electric vector displacesthe free electrons and the columbic electrostaticattraction of the nuclei will restore the electrons tothe original position (figure 1). As a result of theoscillating nature of the electric field of light, theelectron cloud coherently oscillates over the sur-face with a resonance frequency ‘ωp’ [8]. When thefrequency of this oscillation matches with that ofincident radiation, a resonance condition is estab-lished which results in the intense absorption, oftentermed as ‘surface plasmon (SP) absorption’. Forexample, spherical gold nanoparticles exhibit asingle surface plasmon band at around 520 nm,attributed to the collective dipolar oscillation ofthe electron cloud (figure 2).

TUNING FUNCTIONAL PROPERTIES 55

Figure 2. (A) TEM images of gold nanoparticles (obtained by reducing HAuCl4 using gallic acid at room temperature);(B) absorption spectrum of Au nanoparticle and nanorod; (C) TEM images of gold nanorod (obtained by photochemicalreduction); (D) surface plasmon oscillation of spherical gold nanoparticles and (E) the surface plasmon oscillation of goldnanorods in the transverse and longitudinal axes.

For a bulk metal with infinite dimension in allthe three directions, the resonance frequency ofplasma electrons (ωp) can be expressed as

ωp = (Ne2/ε0me)1/2 (1)

where N is the number density of the electrons,ε0 is the permittivity of vacuum, and e and me

are the charge and effective mass of the elec-tron, respectively. Depending on the dimension-ality of the nanostructured materials, differentboundary conditions can be imposed on the elec-tron plasma. In contrast to the bulk metal, the elec-tron cloud in nanoparticles is confined to a finitevolume, which is smaller than the wavelength oflight. Hence, the frequency of oscillation of metalnanoparticles is determined mainly by four para-meters: number density of electrons, effective massof the electron and the size and shape distribu-tion of the charge. This allows the tuning of theoptical properties of noble metal nanoparticles byvarying the size, shape and dielectric environment[8]. In 1908, Gustay Mie provided a quantitativedescription for the resonance in spherical particles,by solving Maxwell’s equations, with appropriateboundary condition [15]. According to Mie theory,the total cross-section consists of scattering andabsorption (often termed extinction) and given assummation over all electric and magnetic oscilla-tions. The contribution of absorption and scatter-ing mainly depends on the size and shape of theparticles.

Nanostructured metallic systems of noble metalsuch as silver, gold and copper are of greatinterest since their localized surface plasmons reso-nate at the visible range of the electromagnetic

radiation. Typically, the surface plasmon (SP)band for spherical Au nanoparticles having dia-meter of ∼20 nm is observed at around 530 nm.The surface plasmon absorption of Au nanopar-ticle obtained by reducing HAuCl4 using gallicacid at room temperature [16] and gold nanorodprepared by photochemical reduction [17] andtheir corresponding TEM images are presented infigure 2. A bathochromic shift in the λmax wasobserved from 530 to 550 nm on increasing thediameter of the nanosphere from 20 to 80 nm,which is attributed to the electromagnetic retar-dation in larger nanoparticles. The dependenceof the nanoparticle diameter on the maximum ofplasmon resonance band was theoretically calcu-lated by adopting Mie theory and discrete dipoleapproximation (DDA) method and these aspectswere reviewed extensively [8]. As the shape ofthe nanoparticle changes, the surface electron den-sity and hence the electric field on the surfacevaries. This causes dramatic variations in the oscil-lation frequency of the electrons, generating dif-ferent optical cross-sections (including absorptionand scattering). For example, nanorods of gold andsilver split the dipolar resonance into two surfaceplasmon bands wherein the induced dipole oscil-lates along the transverse and longitudinal axes(figures 2B and E) [18]. The longitudinal sur-face plasmon band shifts to longer wavelengthswith increase in aspect ratio, while the positionof transverse surface plasmon band remains moreor less unaffected. These results indicate that thedimensionality plays a crucial role in determin-ing the plasmon resonance of metal nanoparti-cles. Au nanoparticles in the size range of 2–5 nmshow interesting quantum size effects and canbehave as conductor, semiconductor or insulator

56 K GEORGE THOMAS

depending on its dimension. The surface plasmonband of Au nanoparticle, which is characteristicof its metallic nature, undergoes broadening anddampening on decreasing the size below 5.0 nm[19]. A sharp decrease in the intensity of surfaceplasmon band is observed for nanoparticles havingdiameter below 3.2 nm due to the transition frommetallic to semiconductor/insulator behavior. Thiseffect is attributed to the onset of quantum sizeeffect and was further established by theoreticalinvestigations. Surface plasmon band is absent forAu nanoparticles having core diameter less than2 nm. The dampening of the SP mode is attributedto the surface scattering of conduction electronsthat follows inverse radius (1/radius) dependence[19,20]. It was found that Au nanoparticles <5 nmin diameter are catalytically active for several reac-tions, for example, catalysts for the oxidation ofCO (an automobile exhaust) at room temperature[21]. Another interesting property of Au nanopar-ticles/nanorods is the high optical cross-sectionof the surface plasmon absorption, which is typi-cally 4–5 orders of magnitude higher than con-ventional dyes [22]. These unique features of Aunanoparticles/nanorods provide excellent opportu-nities for their use in biomedical field for diagnosis,and imaging. The absorbed light in Au nanoparti-cles/nanorods is efficiently converted into localizedheat and this strategy has been successfully usedfor the laser photothermal destruction of cancercells. Au nanoparticles/nanorods can be targetedto tumor site by conjugating them with bioactivemolecules.

One of the most widely investigated nano-structured systems is spherical gold nanoparticlesowing to their (i) stablity, (ii) ease of synthesis,(iii) excellent optical and electronic propertiesand (iv) their ability to bind with thiols, whichallows functionalization with various molecularsystems. In 1857, Michael Faraday provided thefirst systematic investigation and documentationon the synthesis of colloidal Au nanoparticles[23]. An aqueous solution of sodium tetrachloroau-rate, (Na[AuCl4]), was reduced with phospho-rus in carbon disulfide. The solution turned todeep ruby color and Faraday concluded that thegold was dispersed in the liquid in a very finelydivided form. Several synthetic procedures havebeen developed over the years for the design ofAu nanomaterials having varying size and shape.The notable ones include (i) Turkevich methodby boiling an aqueous solution of HAuCl4 andsodium citrate [24], and (ii) two-phase reductionmethod developed by Schiffrin, Brust and cowor-kers [25]. In the last decade we have witnesseda plethora of scientific activities related to thedevelopment of newer synthetic methods, whichled to the size and shape controlled synthesis of

noble metal nanoparticle having tunable optoelec-tonic properties. Control on the shape and size ofmetal nanoparticles have been achieved by vary-ing the reaction conditions such as reduction tech-nique, reaction time and concentration of cappingagent. Recent developments in the classical wetchemistry methods have enabled the synthesis ofanisotropic nanostructures possessing well definedshapes (for example, ellipsoids, rods, cubes, disks,tetrahedra, cylinders, pyramids, triangular prisms,and multipods) and these aspects are discussedextensively [26,27]. The functional properties ofvarious nanomaterials were correlated with theirsize and shape and these aspects are well docu-mented [1–8].

2.2 Semiconductor quantum dots

Photoexcitation of a bulk semiconductor resultsin the transfer of an electron from valence bandto conduction band creating an electron-hole pair,called ‘exciton’, bound by a weak columbic inter-action. The minimum energy required to generatesuch an exciton is called the band gap energy[28]. Excitons can be treated as hydrogen-like sys-tem and the spatial separation between the chargecarrier pair is termed as exciton Bohr radius (aB)which can be deduced from the Bohr approxi-mation. When the physical dimensions of mat-ter become comparable or lower to the excitonBohr diameter (2aB), the functional properties ofa semiconductor becomes sensitive to the size andshape due to confinement of excitons. This pheno-menon is often expressed as quantum size effector quantum confinement effect. For example, CdSehas a Bohr exciton radius of ∼56 A; the elec-tron and hole cannot achieve their desired distancewhen the diameter of nanocrystal is smaller than112 A [29–32]. As mentioned in the previous sec-tion, this situation can be illustrated as a ‘particle-in-a-box’ system and explained from a quantummechanical point of view.

Many theoretical models have been reportedwhich correlate the size of quantum dots to theirband gap energy and these models were furthercompared with the experimentally observed data[34]. An earlier theoretical calculation for semicon-ductor nanoparticles (using CdS and CdSe QDs asexamples) was reported by Brus, based on ‘effec-tive mass approximation’ (EMA) [28–33]. In thisapproximation, an exciton is considered to be con-fined to a spherical volume of the crystallite andthe mass of electron and hole is replaced witheffective masses (me and mh) to define the wavefunction. Kayanuma accounted the electron-holespatial correlation effect and modified the Brusequation [33]. Based on the modified equation, thesize dependence on the band gap energy of QDs

TUNING FUNCTIONAL PROPERTIES 57

Figure 3. (A) Photograph illustrating the size dependent emission from CdSe quantum dots having radius of (a) 2.6 nm(b) 3.7 nm (c) 4.7 nm; (B) the energy levels of the valence and conduction bands of core and shell in type I heterostuctureQDs.

can be quantified as follows:

Eg(QD) = Eg(bulk) + (�2π2/2μR2)

− (1.786e2/εR) − 0.248E∗Ry (2)

where ‘R’ is the radius of QD, μ is the reducedmass, ‘ε’ is the permittivity of the vacuum and E∗

Ry

is the effective Rydberg energy. The first term onthe right hand side represents the band gap of bulkmaterials, which is characteristic of the material:for example, 2.53 eV for CdS, 1.74 eV for CdSe and1.50 eV for CdTe. The second additive term of theequation represents the additional energy due toquantum confinement having a 1/R2 dependenceon band gap energy, Eg(QD). The third subtractiveterm stands for the columbic interaction energyof exciton having 1/R dependence; often neglecteddue to high dielectric constant of the material. Thelast subtractive term, stands for spatial correlationeffect (independent of radius), and significant onlyin case of semiconductor materials with low dielec-tric constant.

The experimental observation of the size depen-dence in the band gap energy of semiconductoris in good agreement with the theoretical predic-tions [35,36]. Variation in the band gap energyof semiconductor with size is directly reflectedon their optical responses (figure 3). The con-finement of electrons in semiconductor QDs influ-ences their electronic structure in two ways. Theenergy gap of semiconductor QDs increases dra-matically with decrease in its radius (Eg ∝ 1/R2).Also the continuum observed in the conductionband and valence band in the case of bulk mate-rials is replaced with discrete atomic like energylevels as the particle size decreases.

Earlier attempts to produce extremely small par-ticles included template-assisted synthesis usingzeolites, micelles, lipid bilayers, molecular sievesand polymers [37]. Most of the initial effortswere concentrated on the synthesis of cadmium

chalcogenide such as CdS, CdSe, CdTe (II–VI semi-conductors). A breakthrough in producing highquality monodispersed QDs of II–VI semiconduc-tors was reported by Bawendi and coworkers in1993 by adopting a high temperature organometal-lic reaction in presence of a coordinating solvent[38]. The reaction was carried out at elevated tem-peratures (∼360◦C) under vacuum, using seleniumprecursor (TOPSe; selenium coordinated to tri-octylphosphine) which was injected rapidly into asolution of dimethylcadmium in TOPO. An excel-lent control over crystal growth was achieved bycontrolling the reaction parameters such as pre-cursor concentration, temperature and duration ofreaction. A marked improvement in the above syn-thesis strategy was achieved by Peng and cowor-kers, where the authors have used a non-pyrophoricand stable cadmium precursor, cadmium oxideinstead of dimethylcadmium [39]. Later severalmodifications have been reported with the aimof improving the monodispersity and increasingthe photoluminescence quantum yield and thesemethods were further extended for the synthesis ofIII–V semiconductor QDs (GaAs, InP, etc.) [40].

A hypsochromic shift in the absorption as wellas emission band was observed on reducing the sizeof semiconductor nanocrystal, which is accompa-nied by the appearance of sharp absorption peaksoriginating from the first and second excitonictransitions [41]. For example, bulk CdSe has anabsorption onset at 720 nm (Eg = 1.74 eV) whichis blue shifted to 450 nm (Eg = 2.8 eV) by reducingthe size down to about 2 nm. Many other propertiesof semiconductor nanomaterials also depend on thephysical dimension of the material. It is reportedthat the redox potential of the valence and con-duction bands are also sensitive to size of the QDs,which shifts to more positive values for valenceband and negative values for conduction band asthe size decreases [42].

Bare QDs possess low emission yield due tothe presence of the dangling bonds and deeptrap states on their surface, which enhances the

58 K GEORGE THOMAS

Figure 4. HRTEM images of (A) CdSe QDs and (B) CdSeQDs overcoated with 3.9 monolayers of ZnS; (C) Absorptionand PL spectra of bare CdSe QDs and overcoated with 3.9monolayers of ZnS (bathochromic shift in absorption bandand enhancement in PL) along with (D) photographs of bare(left flask) and overcoated (right flask) CdSe QDs (repro-duced with permission from [46], copyright 2003 AmericanChemical Society).

nonradiative channels [43]. One of the success-ful strategies adopted for imparting photostabilityand improving the photoluminescence efficiency ofQD is to overcoat with an inorganic shell materialhaving similar lattice parameters and higher bandgap energy (Eg). Such core-shell systems whereinthe core is overcoated with large band gap shellmaterial are called as type I heterostucture QDs(figure 3). In such systems, the conduction band ofthe shell possesses higher energy than that of thecore and the valance band of the shell possesseslower energy than that of the core. As a result,both electrons and holes (exciton) are confinedin the core. Shell materials used for overcoatingCdSe QDs include zinc sulphide (ZnS) [43], zincselenide (ZnSe) [44] and cadmium sulphide (CdS)[45]. Among these, ZnS is widely used as overcoat-ing material owing to its higher band gap energy,allowing efficient confinement of excitons in thecore. Overcoating prevents photoinduced chargetransfer in bare QDs retaining PL efficiency. It alsoprevents the leakage of core materials, avoiding thetoxicity inside of the cell when used in biologi-cal applications. However, the core-shell QDs withlarge shell thickness are less effective for biologicalapplications due to the decrease in luminescencequantum yield. Studies from our group have shownthat two monolayers (corresponding to ∼0.65 nm)of ZnS shell is the optimum shell thickness for a4.2 nm diameter CdSe quantum dot, which inhibitscharge transfer processes and provides maximumPL quantum yield (figure 4) [46].

2.3 Carbon nanotubes

Another class of nanoscale building blocks whichreceived much attention in recent years is thecarbon nanotube (CNT), an allotrope of carbon.They are molecular scale tubes of graphene sheets;an ideal CNT is a hexagonal network of car-bon atoms rolled as a seamless cylinder, withdiameter in the order of few nanometers (0.8–2.0 nm) and a tube length varying in thousandsof nanometer. This class of quasi one-dimensionalnanostructures possess unique electrical propertieswhich make them an attractive candidate for thefabrication of nanodevices and these aspects arereviewed in detail [47–51]. The chemical bondingof nanotubes is similar to that of graphite, com-posed entirely of sp2 hybridized carbon atoms.Carbon nanotubes are mainly classified as twotypes: single-walled carbon nanotubes (SWCNT)and multi-walled carbon nanotubes (MWCNT).Based on the theoretical studies, it was proposedthat the electronic properties of ‘ideal’ carbonnanotubes depend on their diameter and chira-lity. However, experimental studies have showedthat various defects in carbon nanotubes such aspentagons, heptagons, vacancies or dopant candrastically modify the electronic properties ofCNTs [48].

The fascinating electronic properties of carbonnanotubes originates from the quantum confine-ment of electrons normal to the nanotube axis.In the radial direction, electrons are confined bythe monolayer thickness of the graphene sheet andperiodic boundary conditions exist around the cir-cumference of the nanotube. Electrons can onlypropagate along the nanotube axis due to quan-tum confinement and their wave vectors pointin this direction. The resulting number of one-dimensional conduction and valence bands dependson the standing waves around the circumference ofthe nanotube.

The single walled carbon nanotubes are usuallydescribed using the chiral vector, Ch = na1 + ma2,which connects two crystallographically equivalentsites, (A and A′ in figure 5A), on a graphene sheetwhere a1 and a2 are unit vectors of the hexago-nal honeycomb lattice and n and m are integers[48]. When the graphene sheet is rolled up intoa nanotube, the ends of the chiral vector meeteach other to form the circumference of the nan-otube. The chiral vector Ch also defines chiralangle (θ), the angle between the chiral axis and thezigzag axis of the graphene sheet. The various val-ues of n,m and θ, defines the different nanotubestructures such as chiral (n,m), armchair (n, n)and zigzag (n, 0 or 0,m). Armchair nanotubes areformed when n = m having a chiral angle of 30◦

whereas zigzag nanotubes are formed when either

TUNING FUNCTIONAL PROPERTIES 59

Figure 5. (A) Schematic diagram of two-dimensionalgraphene sheet illustrating chiral axis, zigzag axis andchiral angle; (B) HRTEM images of bundled SWCNT;(C) electronic properties of an armchair nanotube exhibit-ing metallic behavior and (D) zigzag nanotube exhibitingsemiconductor behavior (A, C and D are reproduced withpermission from [48], copyright 2002 American ChemicalSociety).

n or m are zero and the chiral angle is 0◦. All othernanotubes with chiral angle between 0◦ and 30◦

are known as chiral nanotubes [48]. The electronicproperties of a nanotube vary in a periodic waybetween metallic and semiconductor by followinga general rule of n − m = 3i, where i is an inte-ger or noninteger. If i is an integer, i.e., (n − m)is a multiple of 3, then the tube exhibits a metal-lic behavior and possesses a finite value of carri-ers in the density of states at the Fermi energylevel (figure 5C). In contrast, if i is a non-integer,then the tube exhibits a semiconducting behaviorand has no charge carriers in the density of states(DOS) at the Fermi energy level (figure 5D). TheDOS of various types of CNTs were experimentallyinvestigated and compared with theoretical models[48]. Although there have been many interestingand successful attempts to grow CNTs by variousmethods, most widely used techniques are (i) arcdischarge, (ii) laser furnace and (iii) chemical vapordeposition (CVD). Commercially available nano-tubes contain a mixture of metallic and semicon-ductor tubes; one third of them are metallic andthe rest are semiconducting. Even though CNTsare considered as versatile building blocks for thedesign of optoelectronic devices, the availability ofcarbon nanotubes in a scalable quantity having(i) uniform diameter and bandgap, and (ii) metal-lic/semiconducting character remains a majorhurdle.

3. Functional properties of hybridnanomaterials

The functional properties of nanomaterials canbe further tuned by the stepwise integrationof nanoscale building blocks (nanoparticles,nanorods, nanotubes, etc.) into hybrid nanoma-terials. When integrated as hybrid nanomaterials,the functional properties of nanoscale buildingblocks may couple each other to yield newer pro-perties. Thus, hybrid nanomaterials may possessnovel properties different from that of isolatedcomponents or possess complementary propertiesuseful for performing specific functions (vide infra).Several strategies have been reported for designinghybrid nanostructures and representative systemsinclude organic-inorganic and inorganic-inorganichybrid systems. Design of organic-inorganic hybridsystems can be achieved by functionalizing photo-or electroactive molecules on to metal or semi-conductor nanomaterials wherein electron/energytransfer processes may occur. Inorganic-inorganichybrid nanomaterials can be obtained through thehierarchical integration of metal and semiconduc-tor building blocks into higher order assemblies.For example, a nanoscale building block ‘X’ canbe coupled using a spacer group with anothernanoscale building block ‘Y’ to yield ‘X-Y’ typehybrid nanostructures (or with similar buildingblock to yield an ‘X-X’ type nanostructure).Properties of these hybrid nanostructures can betuned by varying the length of the spacer group.Design of hybrid nanomaterials with hetero-junction can be achieved by bringing (i) metal andsemiconductor nanoparticles, (ii) dissimilar metalnanoparticles and (iii) dissimilar semiconductornanoparticles. Such hybrid nanomaterials possessproperties that are fundamentally different fromthose of isolated components. Recent studies haveshown that hybrid nanomaterials possess morepromising functional properties than individualbuilding blocks. A brief summary of the recentdevelopments on the design of hybrid nanomate-rials and their functional properties is presentedhere.

3.1 Organic-inorganic hybrid nanomaterials

Stable metal as well as semiconductor nanoparti-cles were prepared by capping with a monolayer oforganic molecules. The ability of functional groupssuch as quaternary ammonium halides, amines,thiols, etc., to bind on to the surface of nanoparti-cles has been exploited for organizing organic mole-cules around them (figure 6). The most widelyused method for the functionalization of organicmolecules on to the surface of metal nanopar-ticles adopts a two-phase synthesis reported by

60 K GEORGE THOMAS

Figure 6. Monolayer-protected clusters (MPCs) having metal core covered with (A) alkyl thiol and (B) monothiol deriva-tive of triethylene glycol as ligand shell; (C) schematic illustration of a core-shell hybrid nanostructure; (D) schematicrepresentation of ligand exchange (place exchange) reaction for preparing mixed monolayers; (E–G) gold nanoparticleshaving mixed monolayers of (E) fullerenethiol and dodecanethiol; (F) pyrenethiol and dodecanethiol; (G) ruthenium tris-bipyridine thiol and monothiol derivative of triethylene glycol.

Brust et al [25] and their modified procedures [52].Such monolayer-protected clusters (MPCs) can bevisualized as three-dimensional assemblies possess-ing a core-shell structure (metal core covered withligand shell; figures 6A–C) [25,52–54]. Solubilityof MPCs can be tuned from nonpolar to polarmedium by varying nature of the stabilizer layer.

Optoelectronic properties of MPCs can be tunedby linking photoactive molecules on to the sur-face of nanoparticles. Several methodologies havebeen developed for incorporating a desired num-ber of chromophores on to the nanoparticle. Forexample, metal nanoparticles bearing mixed mono-layers can be synthesized using ligand exchange(place exchange) reactions reported by Murray andcoworkers [52,53]. The addition of pre-synthesizednanoparticles possessing thiol bearing ligands to asolution of another thiolate molecule having diversefunctionality results in the partial substitution ofthe protecting shell on the nanoparticle surface(figure 6D). This strategy enabled the functionali-zation of a desired number of photoactive or elec-troactive molecules around the metal nanoparticle(figures 6E–G).

Interesting physical processes arise when photo-active or electroactive molecules are linked on tometal nanoparticles [52,55,56]. The most signifi-cant ones are electron/energy transfer processes,which can be further modulated by varying the

size, shape and chemical constituents of nano-materials. It is reported that Au nanoparticlesin the size range of 2–5 nm can behave as con-ductor, semiconductor or insulator depending ontheir dimension due to quantum size effects [19,57].Optoelectronic properties of hybrid nanomaterialscan be tuned by anchoring chromophores on toAu nanoparticles in this size range. Photoinducedelectron transfer process from a chromophoricsystem (pyrene; figure 6F) to Au nanoparticleof ∼2 nm was demonstrated by following tran-sient spectroscopy [56–59]. The effect of the tran-sition behavior of metal nanoparticles was studiedby Dulkeith et al by functionalizing lissaminemolecules on gold nanoparticles of different sizes(1–30 nm) and isolated the resonant energy transferrate from the decay rates of the excited dye mole-cules [60]. The increase in lifetime with decreasein the nanoparticle size was indicative of thedecreased efficiency of energy transfer. It is fur-ther reported that chromophore bound Au/Agnanoparticles possess unique ability to store andshuttle electrons [19,57,61]. A wide variety of chro-mophores have been functionalized on the surfaceof Au nanoparticles (figure 6E–G) and proposedas active components in light harvesting systems.The significant photophysical events occurring inthese hybrid nanomaterials are summarized infigure 7.

TUNING FUNCTIONAL PROPERTIES 61

Figure 7. Excited-state deactivation processes in fluorophore-metal hybrid nanomaterials.

Photoactive molecules linked semiconductorQDs are widely used as hybrid nanomaterialsfor fluorescence resonance energy transfer (FRET)analysis. The use of QD based hybrid nanomateri-als in the detection of biomolecular systems (pro-teins and DNA) have been summarized by Medintzand coworkers [62,63]. Compared to organic fluoro-phores, QDs possess several unique photophysi-cal properties, which make them attractive asbiolabels: (i) broad absorption spectra and largeStokes shifts, (ii) molar extinction coefficients 10to 100 times higher than organic dyes (iii) size-tunable photoluminescent emission with relativelyhigh quantum yield and lifetime (iv) high resis-tance to photobleaching and chemical degradation.The native organic ligands used as a monolayerfor protecting the QDs can be partially exchangedwith a bioactive/photoactive molecule. The emis-sion properties of QDs can be size-tuned to givebetter spectral overlap with a particular acceptordye. It has been shown that the FRET efficiencycan be increased by loading a central QD donorwith multiple fluorophores [63,64]. Unique photo-physical properties of QDs opens up newer oppor-tunities for the design of FRET-based biologicalassays, providing better distance resolution thantraditional donor–acceptor FRET systems. How-ever, the toxicity of QDs is a matter of concern,particularly for biomedical application, and newergeneration QDs such as InP are now considered asalternatives.

3.2 Plasmon coupling in hybrid nanomaterials

The transport of optical energy using materialsthat are considerably smaller than the wavelengthof light is one of most challenging issues in theminiaturization of photonic components (due toproblems associated with the diffraction limit oflight). However, nanostructures can convert pho-tons into surface plasmons that are not diffractionlimited. Within the propagation length, the sur-face plasmon modes can be decoupled to light andthis possibility offers tremendous opportunities

in the design of nanoscale optical and photonicdevices such as metal-nanoparticle based plasmonwaveguides. Design of higher order hybrid nano-materials (for example, one-dimensional arrays ofnoble metal nanoparticles with defined particlespacing) is an essential requirement for achievingthis goal. Lithographic methods such as electronbeam lithography are commonly used for the con-struction of higher order nanostructures and detailsare summarized in recent reviews [65,66]. Maieret al have recently demonstrated the transport ofelectromagnetic energy over a distance of 0.5μmin plasmon waveguides consisting of closely spacedsilver rods [67]. The waveguides were excited by thetip of a near-field scanning optical microscope andenergy transport was probed by using fluorescentnanospheres.

Recent studies have shown that it is possible tofine tune the optical properties of metallic nanopar-ticles by their controlled organization into periodicarrays [65,66]. Two types of interactions exist inorganized metal nanoparticles: near-field couplingand far-field dipolar coupling. Near-field coupling(evanescent coupling) is observed in an ensembleof closely packed nanoparticles wherein they nearlytouch each other. In the latter case, the dipolefield resulting from the plasmon oscillation of metalnanoparticle induces an oscillation in a neighbor-ing nanoparticle.

As mentioned in previous sections, closelypacked 1D arrays of Au nanoparticles can, in prin-ciple, function as (i) guides of electromagneticradiation (waveguides) allowing miniaturization ofdevices below the diffraction limit and (ii) intercon-nectors in optical and photonic devices. However,isotropic nature of spherical Au nanoparticles pre-vents the selective binding of molecules on surfaceswhich restricts the possibility of designing 1D arrayof nanomaterials by chemical functionalizationmethods [56]. In contrast, the anisotropic featuresof Au nanorods allow their assembly in variousorientations, and several attempts have beenmade for organizing Au nanorods using electrosta-tic/supramolecular/covalent approaches [68–73].

62 K GEORGE THOMAS

Figure 8. Molecules used for organizing Au nanorods using hydrogen bonding and electrostatic approaches.

This includes the (i) linear organization of Aunanorods using biotin-streptavidine connectors andlateral organization through electrostatic interac-tions by varying the pH of the medium [72,73],(ii) longitudinal assembly through covalent func-tionalization by using dithiols [68,70], (iii) coop-erative intermolecular hydrogen bonding by using3-mercaptopropionic acid and electrostatic interac-tion by using cysteine and glutathione as shownin figure 8 [69,71] and (iv) end-to-end electrosta-tic assembly of Au nanorods on multiwall carbonnanotubes [74].

On the basis of electron diffraction analysis andHRTEM studies, it is proposed that the end facetsof Au nanorods are dominated by {111} planes andthe side facets by {100} and {110} planes [3,75,76].It is reported that the thiol derivatives preferen-tially bind to the {111} planes of the Au nanorodsand this specific interaction was further exploitedfor the organization of Au nanorods [72]. Thepreferential functionalization at the edges of Aunanorods leads to the formation of 1D nanochainsin the longitudinal direction. Based on detailedmechanistic investigations, it is concluded that thenanochain formation proceeds through an incuba-tion step, followed by the dimerization and subse-quent oligomerization of nanorods in a preferentialend-to-end fashion (figure 9). The clear isosbesticpoint observed in the time dependent absorptionspectrum and dimers observed in the TEM micro-graphs confirms the involvement of the dimeriza-tion step in the chain formation process. Spectralchanges were analyzed for a second-order kineticprocess, and linearity in the initial period furthersupports the dimerization mechanism, which devi-ates with time due to the contribution of othercomplex processes (oligomerization) [70].

More recently, studies are focused on the designof Au nanorod dimers and investigation of plas-mon coupling in these systems as a function of theirdistance and orientation. The plasmon coupling in

Au nanorod dimers linked through rigid molecules(e.g., 1,2-phenylenedimethanethiol) were found tobe more pronounced due to effective dipolar over-lap along their longitudinal axis compared to flex-ible aliphatic dithiol such a 1,6 hexanedithiol [68].These studies confirm that the nature of the linkergroup plays a critical role in plasmon coupling ofmetal nanoparticles. Such hybrid metal nanopar-ticles are promising components in the design ofnanoscale devices, for example stable Au nanorodchains with effective dipolar overlapping can beused as elements in the design of plasmonic waveguides. Thus light waves can be channeled throughhybrid nanomaterials depending on the assem-bly of Au nanorods. Rigid Au nanorod dimersare also promising as nanoscale interconnectors infuture molecular electronics devices and the dis-tance and orientation between the nanorods can bevaried by choosing proper linker groups. A possiblestrategy for the design nanoscale molecular elec-tronic devices using Au nanorod dimers is pre-sented in figure 10.

We have recently developed a novel methodo-logy for the preferential end functionalization of Aunanorods with nanoparticles by exploiting the elec-trostatic attractive interactions [77]. The enhancedpotential at the edges of Au nanorods preferentiallyattracts the positively charged Au nanoparticles,leading to their selective binding. Site specific bind-ing results in a spontaneous bathochromic shiftin the longitudinal plasmon band of Au nanorodswhich is dependent on the size of the nanoparti-cles. This concept can be further utilized for coat-ing dissimilar metals/metal oxides on to the edgesof nanorods and also for alloying specific domainsof nanorods with other metals. Such hybrid mate-rials are useful in surface enhanced spectroscopicstudies. Another possibility is to selectively posi-tion molecules to specific domains of nanomaterialshaving enhanced electric field. This can lead to alarge enhancement of various spectroscopic signals,

TUNING FUNCTIONAL PROPERTIES 63

Figure 9. (A) TEM images of Au nanorods in the absence (a) and presence (b–d) of mercaptoproponic acid (MPA)illustrating the end to end assembly; (B) generalized scheme indicating the stepwise formation of nanochains throughincubation, dimerization and oligomerization; (C) schematic representation of plasmon coupling in dimers and oligomers ofAu nanorod and (D) Absorption spectra of gold nanorods in acetonitrile-water (4 : 1) recorded after the addition of MPA(0–8µM). Decrease in the absorption of the longitudinal plasmon band, accompanied by the formation of coupled plasmonband of Au nanorod dimers, was observed, through an isosbestic point (A and D reproduced with permission from [71],copyright 2004 American Chemical Society and B reproduced with permission from [70], copyright 2006 American ChemicalSociety).

Figure 10. (A) Au dimers as elements in molecular electronics and (B) enhanced potential at the edges of Au nanorods(B reproduced with permission from [77], copyright 2004 American Chemical Society).

finding applications in techniques like surfaceenhanced Raman spectroscopy (SERS) leading tosingle-molecule detection. Design of several ‘meta-materials’ with programmable physical and chemi-cal properties have been reported by Shevchenkoet al through the self-assembly of nanoparticles oftwo different materials into a binary nanoparticlesuperlattice [78].

3.3 Heterojunctions in hybrid nanomaterials

Hybrid nanomaterials possessing heterojunctionsare promising functional nanomaterials havingpotential application in optoelectronic devices.They possess novel properties that are fundamen-tally different from those of nanoscale components[5,79]. Heterojunctions in hybrid nanomaterials

64 K GEORGE THOMAS

Figure 11. (A–C) Various types of heterojunction hybrid nanomaterials of dissimilar semiconductors and (D–E) schematicrepresentation of the spatial separation of excitons in type II core-shell QDs and the corresponding energy levels of thevalence and conduction bands of the core and shell.

can be obtained by bringing nanomaterials of(i) dissimilar metals and (ii) dissimilar semiconduc-tors and (iii) semiconductor and metal.

Heterojunctions of dissimilar metal nanoparti-cles can form alloys having novel optical and elec-tronic properties and these aspects are reviewed[5]. Interestingly when two dissimilar semiconduc-tor nanomaterials are in contact, the heterojunc-tion created can assist the spatial separation ofcharge carriers (excitons) formed upon photoexci-tation. This can be achieved by tuning the energylevels of the valence and conduction bands of semi-conductors. If the valence and conduction bandposition of the semiconductor ‘S1’ is lower (orhigher) than semiconductor ‘S2’, then one carrieris mostly confined in ‘S1’, while the other one in‘S2’ (figure 11). Core-shell nanostructures of thistype are called type-II nanostructures. Such core-shell heterostructures can be synthesized by mole-cular beam epitaxy (e.g., GaSb/GaAs [80]) as wellas chemical methods (e.g., CdTe/CdSe [81,82] andCdSe/ZnTe [81]). Photoexcitation of CdTe/CdSeQDs results in the spatial separation of charge car-riers (excitons); the hole is mostly confined to theCdTe core while the electron is mostly in the CdSeshell. This situation is reversed in CdSe/ZnTe QDs,since the energy levels of the valence and conduc-tion bands of the shell are higher than that ofthe core. In CdSe/ZnTe QDs, the electron residesmostly in the CdSe core, while the hole is mostly inthe ZnTe shell. Emission in type-II nanostructuresoriginates from the radiative recombination of theelectron-hole pair across the core-shell interface.Type-II QDs emit at energies that are smaller thanthe band gap of either material; the photoemission

from CdTe/CdSe and CdSe/ZnTe QDs is observedat longer wavelengths than from the correspondingcore and shell components. The emission propertiesof type-II QDs can be fine-tuned by changing theshell thickness and core size. For example, the emis-sion spectra from CdTe/CdSe QDs can be tuned inthe range from 700 nm to over 1000 nm by chang-ing the core size and shell thickness. Type-II QDsare expected to have longer exciton decay timesdue to the spatial separation of charges; for e.g.,the mean decay lifetimes of CdTe/CdSe hetero-junction QDs is found to be much larger (57 ns)compared to CdTe QDs (9.6 ns). Heterojunctionhybrid nanomaterials of dissimilar semiconductorsare proposed as promising materials for photo-voltaic applications due to the tunability in theemission band and longer lifetimes.

Bulk metals and semiconductors possess differ-ent electrochemical potentials, hence charge redis-tribution occurs at the contact junction so that thepotentials are equilibrated (generally representedas band bending) [83]. When metals are dopedon n-type semiconductor, charge transfer occursto the metal resulting in the depletion of elec-trons at the semiconductor interface. For metalnanoparticle-semiconductor junction, it is difficultto apply traditional Schottky junction behaviordue to issues related to depletion length and afully accepted theoretical model is not yet deve-loped. One of the theoretical models propose thatthe metal nanoparticles may create an interfacial‘Schottky-type’ potential barriers on semiconduct-ing substrates [84]. However, recent experimentalstudies indicate that the Au nanoparticles formsize-dependent ‘nano-Schottky’ potential barriers

TUNING FUNCTIONAL PROPERTIES 65

on semiconducting substrates that asymptoti-cally approach the macroscopic Schottky bar-rier. Hence the semiconductor supported metalnanoparticle will experience this effect and createlocalized depletion regions on the semiconduc-tor wall which act as the deep acceptor states[83]. Other theoretical and experimental inves-tigations dealing with the modified electronicproperties of metal nanoparticle-semiconductorinterface include metal nanoparticle-TiO2 [83],Ag nanoparticle-SbO2 [84], Mn nanoparticle-GaNnanowires [85] and noble metal nanoparticle-SWCNT systems [86]. Methodologies for incorpo-rating Ag, Au and Pt nanoparticles on to thesurface of SWCNT and their potential applica-tions as sensors and field emission transistors [87]have also been demonstrated. It has been recentlyreported that the electron donation to the nanopar-ticle decorated SWCNT network of nanotube fieldemission transistor (NTFET), upon exposure toNO gas, is dependent on the work function of themetal [86]. Based on these studies, it is concludedthat the Schottky type potential barrier existingat the nanoparticle-SWCNT interface is intimatelyrelated to the work function of the metal. WhileSWCNT-metal nanoparticle systems are proposedfor sensing applications, the possibility of utilizingthese materials as components of light energy con-version systems has not been actively pursued.

4. Conclusions and perspectives

Significant progress has been made in this decadeon the synthesis of various nanoscale buildingblocks such as metal and semiconductor nanoma-terials and nanostructured carbon based systems.Ability to design nanoscale building blocks withhigh monodispersity and size and shape controlhas opened up newer opportunities in the fieldof nanoscience and technology. Parallel develop-ments in the field of electron as well as probemicroscopic techniques allowed the characteriza-tion of these systems with atomic scale preci-sion. The functional properties of nanomaterialswere probed using spectroscopic techniques andfurther correlated to their size and shape. Basedon the quantum confinement of electrons, thesize and shape dependent optoelectronic proper-ties of nanomaterials were theoretically predictedand recent experimental studies have confirmedthese aspects. Functional properties of nanomate-rials can be further tuned by the stepwise inte-gration of nanoscale building blocks into hybridsystems. Initial studies have shown that suchhybrid nanomaterials possess newer optical andelectronic properties that are fundamentally differ-ent from those of isolated systems and can serve

as active components in molecular electronics, sen-sors and light energy conversion systems. Designof hybrid nanomaterials with high degree of repro-ducibility and stability is one of the limiting fac-tors preventing their use in devices. Efforts arenow focused on the development of newer strate-gies for the synthesis of hybrid nanomaterials andinvestigation of their novel properties. Studies inthis direction can provide fundamental insight onthe properties and functions of heterojunctions inhybrid nanomaterials.

Acknowledgements

Author thanks the CSIR and JNCASR for sup-port. This is contribution RRLT-PPD-256 fromNIIST, Trivandrum. Special thanks are due toProfessor M V George and Dr Suresh Das (both atNIIST), Professor Prashant V Kamat (Universityof Notre Dame), K Yoosaf and P Pramod (bothformer graduate students at NIIST), R Vinayakan,A R Ramesh, Pratheesh V Nair and Jino George(graduate students at NIIST) for their interestand simulative discussions on this topic at variousstages.

References

[1] Rao C N R, Kulkarni G U, Thomas P J andEdwards P P 2002 Chem. Eur. J. 8 29.

[2] Kreibig U and Vollmer M 1995 Optical properties ofmetal clusters (New York: Springer-Verlag).

[3] Burda C, Chen X, Narayanan R and El-Sayed M A2005 Chem. Rev. 105 1025.

[4] Link S and El-Sayed M A 2000 Int. Rev. Phys. Chem.19 409.

[5] Liz-Marzan L M 2006 Langmuir 22 32.[6] Daniel M-C and Astruc D 2004 Chem. Rev. 104 29.[7] Mulvaney P 1996 Langmuir 12 788.[8] Kelly K L, Coronado E, Zhao L L and Schatz G C 2003

J. Phys. Chem. B 107 668.[9] Schmid G and Corain B 2003 Eur. J. Inorg. Chem.

3081.[10] Alivisatos A P 1996 J. Phys. Chem. 100 13226.[11] Kerker M 1985 J. Colloid Interface Sci. 105 297.[12] Mishra L C 2004 Scientific basis for ayurvedic thera-

pies (CRC Press) 560.[13] Brown C L, Bushell G, Whitehouse M W, Agrawal D S,

Tupe S G, Paknikar K M and Tiekink E R T 2007 GoldBull. 40 245.

[14] Freestone I, Meeks N, Sax M and Higgitt C 2007 GoldBull. 40 270.

[15] Mie G 1908 Ann. Phys. 25 377.[16] Yoosaf K, Ipe B I, Suresh C H and Thomas K G 2007

J. Phys. Chem. C 111 12839.[17] Kim F, Song J H and Yang P 2002 J. Am. Chem. Soc.

124 14316.[18] Link S, El-Sayed M A and Mohamed M B 2005 J. Phys.

Chem. B 109 10531.

66 K GEORGE THOMAS

[19] Templeton A C, Pietron J J, Murray R W andMulvaney P 2000 J. Phys. Chem. B 104 564.

[20] Hostetler M J, Wingate J E, Zhong C-J, Harris J E,Vachet R W, Clark M R, Londono J D, Green S J,Stokes J J, Wignall G D, Glish G L, Porter M D,Evans N D and Murray R W 1998 Langmuir 14 17.

[21] Janssens T V W, Clausen B S, Hvolbæk B, Falsig H,Christensen C H, Bligaard T and Nørskov J K 2007Top. Catal. 44 15.

[22] Jain P K, Huang X, El-Sayed I H and El-Sayed M A2008 Acc. Chem. Res. 41 1578.

[23] Faraday M 1857 Philos. Trans. 147 145.[24] Turkevich J, Stevenson P C and Hillier J 1951 Discuss.

Faraday Soc. 11 55.[25] Brust M, Walker M, Bethell D, Schiffrin D J and

Whyman R 1994 J. Chem. Soc., Chem. Commun. 801.[26] Wiley B, Sun Y G, Mayers B and Xia Y N 2005 Chem.

Eur. J. 11 454.[27] Wiley B, Sun Y G, Chen J Y, Cang H, Li Z Y, Li X D

and Xia Y N 2005 MRS Bulletin 30 356.[28] Scholes G D 2008 ACS Nano 2 523.[29] Brus L E 1983 J. Chem. Phys. 79 5566.[30] Brus L E 1984 J. Chem. Phys. 80 4403.[31] Nirmal M and Brus L E 1999 Acc. Chem. Res. 32 40.[32] Steigerwald M L and Brus L E 1990 Acc. Chem. Res.

23 183.[33] Kayanuma Y 1988 Phys. Rev. B 38 9797.[34] Weller H 1993 Angew. Chem. Int. Ed. 32 41.[35] Wang Y and Herron N 1990 Phys. Rev. B 42 7253.[36] Rama Krishna M V and Friesner R A 1991 J. Chem.

Phys. 95 8309.[37] Pileni M-P 2003 Nat. Mater. 2 145.[38] Murray C, Norris D and Bawendi M 1993 J. Am.

Chem. Soc. 115 8706.[39] Peng Z A and Peng X 2001 J. Am. Chem. Soc. 123

183.[40] Battaglia D and Peng X 2002 Nano Lett. 2 1027.[41] Talapin D V, Rogach A L, Kornowski A, Haase M and

Weller H 2001 Nano Lett. 1 207.[42] Guyot-Sionnest P, Shim M, Matranga C and Hines M

1999 Phys. Rev. B 60 R2181.[43] Dabbousi B O, Rodriguez-Viejo J, Mikulec F V,

Heine J R, Mattoussi H, Ober R, Jensen K F andBawendi M G 1997 J. Phys. Chem. B 101 9463.

[44] Bleuse J, Carayon S and Reiss P 2004 Physica E 21331.

[45] Peng X, Schlamp M C, Kadavanich A V andAlivisatos A P 1997 J. Am. Chem. Soc. 119 7019.

[46] Vinayakan R, Shanmugapriya T, Nair P V,Ramamurthy P and Thomas K G 2007 J. Phys.Chem. C 111 10146.

[47] Iijima S 1991 Nature 354 569.[48] Charlier J-C 2002 Acc. Chem. Res. 35 1063.[49] Ajayan P M 1999 Chem. Rev. 99 1787.[50] Rao C N R and Govindaraj A 2002 Acc. Chem. Res.

35 998.[51] Dresselhaus M S, Dresselhaus G, Saito R and Jorio A

2007 Ann. Rev. Phys. Chem. 58 719.[52] Templeton A C, Wuelfing W P and Murray R W 2000

Acc. Chem. Res. 33 27.[53] Song Y and Murray R W 2002 J. Am. Chem. Soc. 124

7096.[54] Zheng M, Li Z G and Huang X Y 2004 Langmuir 20

4226.[55] Shenhar R and Rotello V M 2003 Acc. Chem. Res. 36

549.

[56] Thomas K G and Kamat P V 2003 Acc. Chem. Res.36 888.

[57] Kamat P V 2002 J. Phys. Chem. B 106 7729.[58] Ipe B I, Thomas K G, Barazzouk S, Hotchandani S

and Kamat P V 2002 J. Phys. Chem. B 106 18.[59] Ipe B I and Thomas K G 2004 J. Phys. Chem. B 108

13265.[60] Dulkeith E, Morteani A C, Niedereichholz T, Klar T A,

Feldmann J, Levi S A, van Veggel F C J M,Reinhoudt D N, Moller M and Gittins D I 2002 Phys.Rev. Lett. 89 203002.

[61] Hirakawa T and Kamat P V 2005 J. Am. Chem. Soc.127 3928.

[62] Sapsford K E, Berti L and Medintz I L 2006 Angew.Chem. Int. Ed. 45 4562.

[63] Medintz I, Uyeda H, Goldman E and Mattoussi H 2005Nat. Mater. 4 435.

[64] Clapp A R, Medintz I L, Mauro J M, Fisher B R,Bawendi M G and Mattoussi H 2004 J. Am. Chem.Soc. 126 301.

[65] Hutter E and Fendler J H 2004 Adv. Mater. 16 1685.[66] Girard C and Dujardin E 2006 J. Opt. A: Pure Appl.

Opt 8 S73.[67] Maier S A, Kik P G, Atwater H A, Meltzer S, Harel E,

Koel B E and Requicha A A G 2003 Nat. Mater. 2 229.[68] Pramod P and Thomas K G 2008 Adv. Mater. 20

4300.[69] Sudeep P K, Joseph S T S and Thomas K G 2005

J. Am. Chem. Soc. 127 6516.[70] Joseph S T S, Ipe B I, Pramod P and Thomas K G

2006 J. Phys. Chem. B 110 150.[71] Thomas K G, Barazzouk S, Ipe B I, Joseph S T S and

Kamat P V 2004 J. Phys. Chem. B 108 13066.[72] Caswell K K, Wilson J N, Bunz U H F and Murphy C J

2003 J. Am. Chem. Soc. 125 13914.[73] Murphy C J, Sau T K, Gole A M, Orendorff C J, Gao J,

Gou L, Hunyadi S E and Li T 2005 J. Phys. Chem. B109 13857.

[74] Correa-Duarte M A, Perez-Juste J, Sanchez-Iglesias A,Giersig M and Liz-Marzan L M 2005 Angew. Chem.,Int. Ed. 44 4375.

[75] Murphy C J, San T K, Gole A M, Orendorff C J,Gao J X, Gou L, Hunyadi S E and Li T 2005 J. Phys.Chem. B 109 13857.

[76] Perez-Juste J, Pastoriza-Santos I, Liz-Marzan L M andMulvaney P 2005 Coord. Chem. Rev. 249 1870.

[77] Pramod P, Joseph S T S and Thomas K G 2007 J. Am.Chem. Soc. 129 6712.

[78] Shevchenko E V, Talapin D V, Kotov N A, O’Brien Sand Murray C B 2006 Nature 439 55.

[79] Cozzoli P D, Pellegrino T and Manna L 2006 Chem.Soc. Rev. 35 1195.

[80] Hatami F, Grundmann M, Ledentsov N N,Heinrichsdorff F, Heitz R, Bohrer J, Bimberg D,Ruvimov S S, Werner P, Ustinov V M, Kop’ev P Sand Alferov Z I 1998 Phys. Rev. B 57 4635.

[81] Kim S, Fisher B, Eisler H J and Bawendi M G 2003J. Am. Chem. Soc. 125 11466.

[82] Yu K, Zaman B, Romanova S, Wang D andRipmeester J A 2005 Small 1 332.

[83] Ioannides T and Verykios X E 1996 J. Catal. 161 560.[84] Chen X H and Moskovits M 2007 Nano Lett. 7 807.[85] Chen X, Lee S J and Moskovits M 2007 Appl. Phys.

Lett. 91 082109.[86] Kauffman D R and Star A 2007 Nano Lett. 7 1863.[87] Star A, Joshi V, Skarupo S, Thomas D and

Gabriel J C P 2006 J. Phys. Chem. B 110 21014.