Journal of Colloid and Interface Science 255, 79–90 (2002)doi:10.1006/jcis.2002.8558

Formation of Cu and Cu2O Nanoparticles by Variation of the SurfaceLigand: Preparation, Structure, and Insulating-to-Metallic Transition

Mohammed Aslam,∗ G. Gopakumar,∗ T. L. Shoba,∗ I. S. Mulla,∗ K. Vijayamohanan,∗,1

S. K. Kulkarni,† J. Urban,‡ and W. Vogel‡∗Physical and Materials Chemistry Division, National Chemical Laboratory, Dr. Homi Bhabha Road, Pune, India; †Department of Physics, University of Pune,

Pune, India; and ‡Abt. Anorganisch Chemie, Fritz-Haber-Institut der Max-Planck-Gesellscchchaft, Germany

Received July 25, 2001; accepted June 26, 2002

Copper and copper (I) oxide nanoparticles protected by self-assembled monolayers of thiol, carboxyl, and amine functionalities[X(CH2)n–CH3, where X can be –COOH, –NH2, or –SH] have beenprepared by the controlled reduction of aqueous copper salts usingBrust synthesis. The optical absorption spectrum (λmax = 289 nm) isfound to be invariant with the nature of the capping molecule whilethe particle shape and distribution are found to depend stronglyon it. A comparison of the protection efficiency for different cap-ping agents such as dodecanethiol (DDT), tridecylamine (TDA),and lauric acid (LA) suggests that although zerovalent Cu is ini-tially formed for dodecanethiol, all other cases allow oxidation toCu2O nanoparticles. Despite the variation in particle size and rela-tive stability, nanoparticles have been found to form oxides after afew days, especially for the case of LA and TDA surface capping.For all the samples studied, the size has been found to be 4–8 nmby high-resolution transmission electron microscopy. The protec-tive ability is found to be better for dodecanethiol SAM (similar tothe case of Au and Ag nanoparticles), while the order of cappingeffeciency varies as Cu–DDT > Cu–TDA > Cu–LA. In the presentstudy we also demonstrate a reversible metal–insulator transition(MIT) in capped nanoparticles of Cu using temperature-dependentelectrical resistivity measurement. However, the LA-capped sam-ple does not show any such transition, possibly due to the oxideformation. C© 2002 Elsevier Science (USA)

INTRODUCTION

Self-assembled monolayers are increasingly exploited asa method of stabilizing selected metal nanoparticles so thattheir size-dependent electronic properties and chemical reactiv-ities can be conveniently investigated (1–4). These monolayer-protected nanoparticles (MPCs) are useful for several potentialapplications including the design of multifunctional catalysts(5), chemical sensors (6), nucleation control in templated syn-thesis (7), and circuit components such as single electron tran-sistors (SET) in molecular electronics (8). More significantly,since nanoparticles provide a bridge between the properties of

1 To whom correspondence should be addressed. Fax: 091-020-5893044.E-mail: viji@ems.ncl.res.in.

7

metal atoms and those of infinite metallic solids, this way ofexperimental preparation of protected nanoparticles allows di-rect comparison of theory and experiment, notwithstanding theshape distortion and surface state formation of real nanoparticlesdue to the presence of capping functional groups (9).

Among thiol-capped noble metal nanoparticles, gold and sil-ver probably are studied extensively due to their simplicity ofpreparation and relative stability (2, 10). More specifically, dif-ferent types of nanocluster preparations of Au and Ag based onBrust synthesis have given a variety of size-selected nanoparti-cles with unique optical and electronic behavior. In comparison,very few reports (11, 12) are available about the synthesis, sta-bility, and detailed characterization of capped copper nanoparti-cles, although copper nanoparticles are well studied theoreticallywith respect to the size evolution of properties such as ionizationpotential, electron affinity, and band gap (13). The most signifi-cant work is by Pileni et al. (14–17), where the reverse micellarsynthesis route has been used to obtain copper nanoparticles/nanorods of controlled size (2–12 nm) and shape. However, thestability of the nanoparticles is adversely affected by the watercontent leading to copper oxide layer formation after certainlimits (14–17). One reason is perhaps the defective quality ofself-assembled monolayers on copper due to the hindrance ofchemisorption of thiols, disulfides, and similar molecules bytheir facile oxidation (18). However, recently Rubinstein andco-workers have reported (19) superior SAM formation on oxi-dized metal surfaces and hence capped nanoparticles of coppermay be synthesized if suitable capping molecules and appro-priate preparation conditions are selected. This is further sup-ported by the recent results on SAM formation of Jennings andLaibinis on underpotentially deposited copper substrates wherehighly organized monolayers, robust in liquid and vacuum envi-ronments, have been formed using different thiols (20). Hencestable Cu nanoparticles could be synthesized by protecting themin situ using organic molecules with -SH and other head groupscapable of monolayer formation.

The present work describes the synthesis and characteri-zation of monolayer-protected Cu nanoparticles (MPCs) us-ing different capping agents. We show that the selection ofcapping agents with different end-functional entities such as

9 0021-9797/02 $35.00C© 2002 Elsevier Science (USA)

All rights reserved.

80 ASLAM

1-dodecanethiol, tridecylamine, and lauric acid plays a criti-cal role in controlling stability and ease of formation duringBurst synthesis. The nanoparticles prepared using these cappingagents were characterized by microanalysis, UV–visible spec-troscopy, X-ray diffraction (XRD), X-ray photoelectron spec-troscopy (XPS), and high-resolution transmission electron mi-croscopy (HRTEM). This study complements a recent report onthe preparation of zerovalent copper nanoparticles in aqueousand methanolic solutions using poly(amidoamine) dendrimers(21); however, the focus of that study was synthesis of nanocom-posites after complexation with various ligands followed by op-tical characterization. In contrast to that study, we focus on thepreparation and characterization of different types of coppernanoparticles using self-assembled monolayers of various func-tional groups and the relationship between monolayer structureand cluster stability. In the present study, we also report a novelobservation of the reversible transition from insulating to metal-lic behavior at low temperatures for protected nanocluster arraysusing temperature-dependent electrical resistivity measurement.The disappearance of the Kubo gap at low temperatures in thesesystems, where the interparticle spacing (0.5–1 nm) is less thanthe nanocluster dimensions (5–10 nm), is explained to affect thetransition due to strongly coupled charge fluctuations and clustervibrations.

EXPERIMENTAL

Materials

CuCl2 · 2H2O, NaBH4, 1-dodecanethiol, tridecylamine, andlauric acid (99.9%; Aldrich) were used as received. All otherreagents for the experimental work were obtained from standardsources and were used without further purification.

Cluster Synthesis

The synthesis followed a standard procedure within whichthree conditions were strictly followed: (1) The molar ratio ofthe copper ion to the capping agent was kept constant; (2) Theexperiments were performed in an ice bath; (3) The addition ofNaBH4 was kept at a constant slow rate. For example, 60 mlof 2.5 mM aqueous copper chloride solution was mixed with200 ml of toluene containing 1 mM of capping agents followedby vigorous stirring. To this, 50 ml of 0.4 M NaBH4 aqueoussolution was slowly added to obtain a black solution. The stirringwas continued for 8 h and was stopped when the black aqueouslayer became clear, indicating that most of the Cu ions weretransferred into the nonaqueous layer. After the organic layerwas separated using a dry clean separating funnel, the samplewas carefully dried below 80◦C to obtain a dark brown powder.

Spectroscopic Characterization

UV–Vis Spectroscopy

The optical absorption spectra were recorded using aSchimadzu UV-2101 PC spectrophotometer with spectral

ET AL.

resolution 2 nm. The dried samples were redissolved in drytoluene (25 mg of dried nanoparticles in 50 ml of toluene) beforethe spectra were taken.

X-Ray Diffraction

The powder X-ray diffraction data of the nanocluster wasacquired on a Guinier diffractometer (Huber) equipped with arotating anode at λ = 1.5404 A (CuKα) radiation. The originalpowder was pressed into a thin cubic pellet 8 × 15 × 0.3 mmin size, and the diffraction patterns taken were with the diffrac-tometer set to the −45◦ transmission position. The patterns werecorrected for background scattering and for the usual angle-dependent correction factors (polarization, absorption, geome-tric factor).

X-Ray Photoelectron Spectroscopy

X-ray photoelectron spectroscopy (XPS) measurements wereperformed using a VG Scientific ESCA LAB Mk II spectrometeroperating at better than 10−9 Torr using a monochromatic MgKα

source (hν = 1253.6 eV). The alignment of binding energy wasdone using the C1s binding energy of 285 eV as a reference.The X-ray flux (power 70 W) was kept deliberately low to re-duce the beam-induced damage (22). Acquisition took about15 min. for each core level. All regions, C1s, S2p, O1s, N1s,and Cu2p, were measured with 100 eV pass energy. A thick filmof the samples (spread on nickel sample stubs) were used for themeasurements.

High-Resolution Transmission Electron Microscopy

The TEM experiments were performed on a Philips CM200FEG microscope equipped with a field emission gun with an ac-celerating voltage of 200 kV. The magnification was 389,000×and the coefficient of spherical aberration was 1.35 mm. The im-ages were digitized at a size of 256 × 256 pixels with a pixel sizeof 0.03994 nm. Atomically resolved images were thus possible.Images were stored in a computer after digitization and fur-ther processed. Power spectra were calculated so that structuralanalysis such as interplaner distances and angle between planescould be determined. A drop of liquid containing nanoparticleswas placed on an amorphous carbon film (3 nm thick) depositedon a copper grid. After the evaporation of the liquid, the gridwas introduced into the electron microscope and images as wellas power spectra were recorded.

RESULTS AND DISCUSSION

Elemental Analysis

Elemental analysis data (65.2% Cu, 24.54% C, 3.74% H,and 6.52% S) for Cu-DDT suggests a Cu : S molar ratio of ap-prox. 4 : 1, confirming a compact monolayer similar to that re-ported earlier for thiol-capped Au particles (23). The C : H, C : S,

and S : H ratios are also similar to that for pure dodenethiol,within the experimental uncertainties. The LA-capped copper

Cu AND Cu2O NA

nanoparticles exhibit similar microanalysis data (55.3% Cu,32.64% C, 4.76% H, and 7.25% O), confirming monolayer for-mation on the cluster surface. In comparison, the TDA-cappedCu MPC gives an elemental analysis as 81.65% Cu, 13.2% C,4.29% H, and 0.86% N, in agreement with the correspondingCu/N molar ratio of 7.5 : 1 reported for Au nanoparticles (24),revealing that the amines form less densely packed monolayersthan the corresponding alkanethiols. The ratios C : H, C : N, andN : H are also similar to those for neat tridecylamine, within ex-perimental uncertainties. Thermogravimetric analysis has alsobeen performed to confirm the adsorption of these organicmolecules onto the cluster surfaces. The weight loss (approx.30%) at 270◦C (B.P. for DDT) is in agreement with that of themicroanalysis results. In comparison, TGA- and DTA-cappednanoparticles show weight loss at a slightly lower temperature(approx. 240–250◦C), confirming the desorption of less stableamine and carboxyl groups on the Cu/cuprite cluster surface.

UV–Vis Spectroscopy

Figure 1 shows the superimposed optical absorption spec-tra of Cu nanoparticles protected with different capping agents,1-dodecanethiol (a), tridecylamine (b), and lauric acid (c), intoluene medium. All the spectra display a surface plasmon peakat a very short wavelength, i.e., ca. 289 nm for different cappednanoparticles, suggesting the presence of very small separatedCu nanoparticles (≤4 nm in size). Balogh and Tomalia haveearlier reported (21) such a short-wavelength plasmon for Cunanoparticles, while according to Pileni and co-workers (25)the surface plasmon band should appear around 570 nm if theparticle size is larger than 4 nm. Interestingly, dodecanethiol-protected nanoparticles show a sharp intense plasmon peak whilelauric acid-capped nanoparticles show the lowest intensity peakat the same wavelength, perhaps due to the extent of delocal-ization caused by size variation. Also, the decrease in the ab-sorbance indicates a decrease in the copper particle quantity,revealing the possibility of oxide formation with time on thecopper particle surface (26). Another additional feature, promi-nent in these short wavelength plasmon peaks is the presence ofa shoulder at lower wavelength arising due to the flocculation ofthe Cu nanoparticles after dispersion in the solvent.

Conductivity Measurements

One of the crucial aspects of the above nanoparticles is thatthey have size 5–10 nm and are interlinked by the fragile organicthiol molecules, where the length of the molecule is small com-pared to r , the size of the nanoparticles (please see the TEM re-sults discussed below). The precise location of the nanoparticlescannot be fixed due to the orientational and conformational flex-ibility of the organic matrix and also due to the elastic moduli ofthe matrix, reported to be three orders of magnitude lower thanthose of normal solids (27). According to Gorelik et al. (28),

these type of nanoparticles can vibrate with typical frequenciesof 109 to 1010 Hz and when we compare this with the normal RC

NOPARTICLES 81

FIG. 1. Superimposed optical absorption spectra of Cu nanoparticles pro-tected with different capping agents of 1-dodecanethiol (a), tridecylamine(b), and lauric acid (c) in toluene medium.

time constants of these nanoparticles (10−10–10−11 s) it is clearthat vibrations can be coupled with charge fluctuations. This canbe controlled by changing the organic molecule (both length andthe functionality) and hence we have selected three different cap-ping agents for a representative example of Cu nanoparticles.

The ρ observed for Cu-DDT is on the order of 2 M cm,revealing the insulating nature of this monolayer (Fig. 2). At272 K, ρ decreases sharply and the array goes to the metallic

FIG. 2. Superimposed resistivity vs temperature (K) plots for coppernanoparticles with different capping agents. Cu-LA indicates copper nanopar-

ticles capped with lauric acid, Cu-DDT indicates dodecanethiol-capped coppernanoparticles, and Cu-TDA indicates tridecylamine-capped nanoparticles.

82 ASLAM

state, with a typical ρ value of 20 cm. Perhaps, as the tempera-ture decreases, the hopping of the electron enhances the conduc-tivity in these organically capped nanoparticles. In comparison,the resistivity of TDA-capped nanoparticles shows a transitionat 255 K, while LA-cappped nanoparticles show a gradual de-crease in ρ and remain in the M range up to 25 K, indicatingthe ordering of monolayers. Although all the molecules havelength about 1.9 nm, this behavior may result from changes in theorientation of the organic molecule, which can alter the extentof disorder and hence the coupling of the nanoparticles. Thisvariation of Tc (Cu-DDT > Cu-TDA) can be explained on the

basis of compactness and defects in the capping layer as theinsulating nature and capping efficiency decrease in the order

8 ± 2 nm for DDT- and TDA-capped nanoparticles, which isin excellent agreement with the experimentally observed size.

FIG. 3. (a) The XP spectra of copper nanoparticles, where Cu2p3/2 and Cu2p1/

to Cu-DDT indicating two peaks for sulfur core level spectra; (c) XP spectrum of

ET AL.

DDT > TDA > LA for similar-sized protected nanoparticles. Inorder to determine the nature of the electron transport, we usedvarious thermally activated hopping models (29) in the initial re-gion prior to Tc. For example, the application of ρ ∝ exp(E/kT )and its fitting to the Motts variable range-hopping (VRH) model,ρ ∝ exp (To/T )φ , where To = β/kB N (E) ξ d (30), gives a hop-ping distance comparable to the lattice spacing. The value of φ,related to the dimensionality of the system, is close to 0.5 forboth DDT- and TDA-capped nanoparticles, while an unusuallylow (0.18) value is obtained for the LA-capped nanoparticles.The corresponding localization radius calculation gives about

2 signals appear at 932.4 and 952.3 eV, respectively; (b) sulfur peak correspondingnitrogen peak for the TDA-capped Cu nanoparticles.

Cu AND Cu2O NA

In comparison, the LA-capped nanoparticles show a localiza-tion length of only 4 nm, perhaps due to the possibility ofspatially segregated oxide formation on the cluster surface asthe proximity of the carboxyl group with two oxygen atomsmay cause hydrogen bonding. This possibility, along with theabsence of a sharp transition, suggests that charge fluctua-tions are not very effective in these nanoparticles. More de-tailed studies of the Tc as a function of size with the samecapping molecule should yield a better understanding of thesystem.

X-Ray Photoelectron Spectroscopy (XPS)

Figure 3a shows the XP spectra of Cu 2p3/2 and Cu 2p1/2,which appear at 932.4 and 952.3 eV. The binding energies ofthe corresponding elements are calculated with the reference ofC binding energy, 285 eV. A high-resolution spectrum between925 and 960 eV indicates a doublet with a peak separation of19.9 eV owing to the presence of the Cu zero state since thepeak position, lineshape, and peak-to-peak separation of the Cudoublet is a standard measure of the Cu oxidation state. Theintensity ratio is found to be 1 : 2 for this pair. The Gaussiandeconvolution of the Cu2p3/2 peak due to the presence of ashoulder at a higher BE value shows a peak at 932.6 (±0.1 eV)with a difference of 0.2 eV from the Cu peak. The peak maybe due to the Cu2O present in the sample, the concentration ofwhich increases due to X-ray beam damage. In Cu-TDA andCu-LA nanoparticles, deconvolution of these peaks shows twoGaussian pairs corresponding to Cu2p3/2 and the Cu2O peak.The shift also indicates that the cluster has undergone beam-induced damage as the intensity of the Cu2p3/2 peak has beenconsiderably decreased. The formation of Cu2O nanoparticlescan also be compared with the earlier reported XPS spectra ofCu surfaces showing superior monolayer formation on oxidizedsurfaces (19). The Cu-DDT nanoparticles were found to haveB.E. similar to that of the monolayer-functionalized Cu (2D)surfaces (19). However, for the Cu-TDA and LA capping agentsthe spectra resemble that for the aged samples (19), confirm-ing the better stability of the dodecanethiol monolayer on clus-ter surface. The presence of a shake-up feature as compared tothe earlier reported data also confirms the possibility of someamount of metal hydroxide on the cluster surface. As the B.E.sof metallic copper and Cu(I) are not easily distinguishable, theAuger Cu(LMM) spectrum was also examined; we observeda clear difference between the samples: Cu-DDT showed arelatively strong signal for Cu and a weak signal for Cu2O,while the reverse was obtained for LA- and TDA-capped nano-particles.

Figure 3b shows x-ray photoelectron spectroscopic analysisof the sulfur peak corresponding to Cu-DDT, indicating twopeaks for sulfur core-level spectra. The peak at lower bindingenergy after deconvolution shows two Gaussian pairs consistingof spin–orbit components separated by a difference of 1.1 eV.

The least-squares fit shows that there are two species of sulfurseparated by 4.4 eV. The lower peak, ∼163 eV, is in good agree-

NOPARTICLES 83

ment with that reported earlier (31) for chemisorbed thiol. Thesulfur peak at ∼167 eV is due to a sulfate or sulfonic acid moiety(31) formed during the beam-induced damage of the monolay-ers. Figure 3c shows the XP spectrum of the nitrogen peak forthe TDA-capped Cu nanoparticles. The 1s1/2 peak for nitrogenappears at 399.8 eV. The decrease of 0.2 eV from the standardvalue of 400 eV indicates that it forms a stable protective mono-layer on the Cu surface.

X-Ray Diffraction Analysis (XRD)

Figure 4 shows the comparison of XRD patterns of(a) dodecanethiol-, (b) tridecylamine-, and (c) lauric acid-cappedcopper nanoparticles obtained after the nanoparticles are dried atambient conditions. Both metallic copper and cuprite phases canbe identified and this clearly shows that the zerovalent coppernanoparticles formed in the chemical reduction stage undergodecomposition due to limited stability after solvent removal. Ascompared to Cu-TDA and Cu-LA, the XRD pattern of Cu-DDTshows characteristic low-angle peaks corresponding to largerunit cells. Such a lattice composed of alkanethiolate-protectednanoparticles can only result from the periodic arrangement ofthe nanoparticles. The peaks at 0.48 and 0.78 confirm the pres-ence of metallic copper, which is similar to that reported earlier

FIG. 4. The comparison of XRD patterns of (a) dodecanethiol-,

(b) tridecylamine-, and (c) lauric acid-capped copper nanoparticles obtainedafter drying the nanoparticles at ambient conditions.

84 ASLAM ET AL.

FIG. 5. TEM images of (a) dodecanethiol-, (b) tridecylamine-, and (c) lauric acid-capped copper nanoparticles after drying and redispersal in toluene. The

insets are the histograms with the average diameter of the copper nanoparticles with corresponding capping agents. The films were drop-caste from the toluene solutions onto carbon-coated Formvar films on Cu grids.

for the nanoparticles synthesized using laser irradiation (32). Acalculation of the particle size by the Scherrer formula gives the

average diameter as 10–12 nm, which is almost double that ob-tained from the UV results, indicating that solvent removal has

caused agglomeration of the nanoparticles. The relative inten-sity ratio of the major peaks of Cu2O(111) shows that Cu-DDT

has a smaller proportion of oxide and in dodecanethiol-cappednanoparticles Cu : Cu2O is 1 : 1 while TDA and LA contain

85

TEM images of atoluene are shown in

Cu AND Cu2O NANOPARTICLES

FIG. 5—Continued

approx. 25% of copper in the metallic state. Thus the functionalgroups on the alkyl chains on the surface of the cluster seem toplay a very important role in controlling the transformation ofzerovalent copper to cupric oxide.

High-Resolution Transmission ElectronMicroscopy (HRTEM)

ll samples after drying and redispersing inFig. 5 in lower resolution mode. From the

TE micrograph of dodecanethiol-capped nanoparticles (Fig. 5a),it appears that the particles are spheroid in shape with anaverage diameter of approx. 5–7 nm. The histogram for theparticle size distribution shows uniformity. In sharp contrast,tridecylamine-capped nanoparticles are found to be crosslin-ked to form tubular clumps with diffuse contours, attributedto the presence of amorphous particles; the particle size distri-bution (histogram) gives about 6 nm (standard deviation 10%)size and interestingly the short distance between the nano-

particles in large aggregates could be explained using the in-teraction of the nonpolar part of the capping molecule. In

86

spacon

ASLAM ET AL.

FIG. 5—Continued

comparison, the LA-capped nanoparticles show agglomeratesof size around 5–7 nm at lower resolution but with a hexagonalfaceting.

Figure 6 shows the high-resolution TEM images fordodecanethiol- (a), tridecylamine- (b), and lauric acid-(c) capped copper nanoparticles, illustrating the crystallinity ofall the samples; the faceting in the grain is also seen clearly.In the close-up of the nanoparticles, parallel lines with an in-terplanar spacing of 2.82 A can be seen, corresponding to thelateral projection of (111) planes of Cu2O; also the interplanar

cing for LA-capped nanoparticles is found to be 2.4 A. Intrasts tridecylamine-capped nanoparticles show a spacing of

1.08 A, indicating the possibility of compression in the lattice.This lateral fringe is seen uniformly over the entire interior ofthe nanoparticles.

The inset of Fig. 6a shows a higher resolution image of oneparticle. These studies confirm the assignment to the cubic phaseand enable us to ascertain the particle size distribution. Mostof the nanoparticles are cubic in nature (inset) and one of theinsets shows that half of the particle is decahedrally orientedalong the 5-fold axis. The inset of Fig. 6b is also shown toindicate cubic morphology, while in the case of Cu-LA-capped

nanoparticles, the fcc nature is more evident in the inset withhexagonal faceting.

A

Cu AND Cu2O N

A combined analysis of the above data (Table 1) shows thatthe surface capping agent plays an important role in controlling

the particle size, shape, and stability. Although UV–vis spectro-scopic studies show that the surface plasmon band position is in-

suggest a special role played by the monolayers on the surface instabilizing the nanoparticles while TG/DTA results illustrate the

FIG. 6. The high-resolution TEM images for dodecanethiol- (a), tr

NOPARTICLES 87

dependent of the specific ligand, all other studies suggest distinctdifferences between the nanoparticles. For example, IR studies

idecylamine- (b), and lauric acid- (c) capped copper nanoparticles.

88 ASLAM

TABLE 1

Cluster size Decomposition Shape Percent(σ = ±5%) temperature of the composition

Capping agent (using TEM) (nm) TDECa (K) particles Cu : Cu2O

Dodecaethiol 5 nm 550 Spherical 1 : 1Tridecylamine 6.5 nm 445 Tubular 1 : 3.5Lauric acid ∼10 nm 508 Hexagonal 1 : 3.5

faceting

a Temperature in TGA experiments corresponding to a 45% weight loss.

n-

ferent “functional group–metal” bonding energies and enviroments.

FIG. 6—Co

ET AL.

variation of thermal stability, as the LA-capped nanoparticles areless stable than their thiolate counterpart. HRTEM studies con-firm the variation in core size for different capping agents, whichis generally not observed in alkanethiolate–Au-MPCs, where thecore changes little with chain length. More significantly, our ex-perimental results show the dodecanethiol-capped nanoparticlesto be spherical, while TDA- and LA-capped nanoparticles seemto have tubular and hexagonally truncated shapes, respectively.The trends in the stability vary as –SH > –NH2 > –COOH andthe variation in shape and size with different functional groupsbut similar chain length could be explained on the basis of dif-

ntinued

89

ers.ity,

Cu AND Cu2O NANOPARTICLES

FIG. 6—Co

CONCLUSIONS

In summary, copper nanoparticles of different size, shape, andsurface capping have been synthesized with DDT, TDA, andLA as the organic monolayers for protection. Cu nanoparticlescapped with DDT are found to be relatively stable. XRD andXPS results confirm that the Cu nanoparticles synthesized usingthe Brust synthesis route have very limited stability under am-bient conditions, especially for TDA and LA capping agents, asthey form cuprite nanoparticles despite retaining the monolay-

The possibility of controlling the surface properties, stabil-and reactivity of particles chemically using different capping

ntinued

agents provides the opportunity to correlate shape and propertieswith functional groups on the surface. UV–vis and HRTEM haveshown the particles to be in the range of 4–7 nm although thestability varies drastically with respect to the nature of the cap-ping molecule. Disorder-induced, reversible insulator-to-metaltransitions in these quantum dots reveal that the interparticleseparation is much smaller than the size of the nanoparticles.

ACKNOWLEDGMENTS

Mohammed Aslam thanks the Council of Scientific and Industrial Researchfor the award of a senior research fellowship. The authors thank Dr. A. Bhaskaran

90 ASLAM

for allowing us to do the UV–vis spectroscopy, Dr. A. B. Mandale for XPS, andDr. C. Gopinathan for FT-IR.

REFERENCES

1. Templeton, A. C., Wuelfing, W. P., and Murray, R. W., Acc. Chem. Res. 33,27 (2000).

2. Lisiecki, I., Sack-Kongehl, H., Weiss, K., Urban, J., and Pileni, M. P.,Langmuir 16, 8807 (2000).

3. Sun, S. H., Murray, C. B., Weller, D., Folks, L., and Moser, A., Science287, 1989 (2000).

4. Ahmadi, T. S., Wang, Z. L., Green, T. C., Henglein, A., and El-Sayed, M. A.,Science 272, 1924 (1996).

5. Tomalia, D. A., and Dvornic, P. R., Nature 372, 617 (1994).6. Elghanian, R., Storhoff, J. J., Mucic, R. C., Letsinger, R. L., and Mirkin,

C. A., Science 277, 1078 (1997).7. Tanori, J., and Pileni, M. P., Langmuir 13, 639 (1997).8. Feldiem, D. L., and Keating, C. D., Chem. Soc. Rev. 27, 1 (1998).9. Bawendi, M. G., Steigerwald, M. L., and Brus, L. E., Annu. Rev. Phys.

Chem. 41, 477 (1990).10. Henglein, A., Chem. Rev. 89, 1861 (1989).11. Petit, C., Lixon, P., and Pileni, M. P., J. Phys. Chem. B 97, 12974

(1993).12. Bagwe, R. P., and Khilare, K. C., Langmuir 13, 6432 (1997).13. Crispin, X., Bureau, C., Geskin, V., Lazzaroni, R., and Bredas, J. L., Eur.

J. Inorg. Chem. 99, 349 (1999).

14. Tanori, J., and Pileni, M. P., Adv. Mater. 7, 862 (1995).15. Tanori, J., and Pileni, M. P., Langmuir 13, 639 (1997).

ET AL.

16. Pileni, M. P., Gulik-Krzywicki, T., Tanori, J., Filankembo, A., and Dedieu,J. C., Langmuir 22, 7359 (1998).

17. Lisiecki, I., Filankembo, A., Sack-Kongehl, H., Weiss, K., Pileni, M. P.,and Urban, J., Phy. Rev. B 61, 4968 (2000).

18. Laibinis, P. E., and Whitesides, G. M., J. Am. Chem. Soc. 114, 1990 (1992).19. Ron, H., Cohen, H., Matlis, S., Rappaport, M., and Rubinstein, I., J. Phy.

Chem. B 102, 9861 (1998).20. Jennings, G. K., and Laibins, P. E., J. Am. Chem. Soc. 119, 5208 (1997).21. Balogh, L., and Tomalia, D. A., J. Am. Chem. Soc. 120, 7355 (1998).22. Colvin, V. L., Goldstein, A. N., and Alivisatos, A. P., J. Am. Chem. Soc.

114, 5221 (1992).23. Leff, D. V., Ohara, P. C., Heath, J. R., and Gelbart, W. M., J. Phy. Chem.

99, 7036 (1995).24. Leff, D. V., Brandt, L., and Heath, J. R., Langmuir 12, 4723 (1996).25. Lisiecki, I., Billoudet, F., and Pileni, M. P., J. Phys. Chem. 100, 4160 (1996).26. Salkar, R. A., Jeevanandam, P., Kataby, G., Aruna, S. T., Koltypin, Y.,

Palchik, O., and Gedanken, A., J. Phys. Chem. B 104, 893 (2000).27. Polymeropoulos, E. E., and Sagiv, J., J. Chem. Phy. 69(5), 1836 (1978).28. Gorelik, L. Y., Isacsson, A., Voinova, M. V., Kasemo, B., Shekhter, R. I.,

and Johnson, M., Phys. Rev. Lett. 80(20), 4526 (1998).29. Mott, N. F., and Davis, E. A., “Electronic Processes in Noncrystalline Ma-

terials.” Clarendon, Oxford, 1971.30. We have calculated the localization radius using, To = β/kB N (E) ξd , where

N (E) is the density of states at Fermi level, ξ the localization radius ofstates near the Fermi level, d the dimensionality (considered as 3), and β anumerical constant (21.2 for 3D) characteristic of the system.

31. Bourg, M. C., Badia, A., and Lennox, R. B., J. Phys. Chem. B 104, 6562(2000).

32. Yeh, M., Yang, Y., Lee, Y., Lee, H., Yeh, Y., and Yeh, C., J. Phys. Chem. B103, 6851 (1999).