Versatile Synthesis

Versatile Synthesis of Individual Single-Walled Carbon Nanotubes from Nickel Nanoparticles for the Study of Their Physical Properties

Matthieu Paillet,*Vincent Jourdain,Philippe Poncharal,Jean-Louis Sauvajol,and

Ahmed Zahab

Groupe de Dynamique des Phases Condense′es https://www.360docs.net/doc/7815576215.html,RS5581,CC026-Uni V ersite′Montpellier II,

34095Montpellier Cedex5,France

Jannik C.Meyer and Siegmar Roth

Max-Planck-Institut fu¨r Festko¨rperforschung,Heisenbergstrasse1,70569Stuttgart,Germany

Nade`ge Cordente,Catherine Amiens,and Bruno Chaudret

Laboratoire de Chimie de Coordination du CNRS,205,route de Narbonne,31077Toulouse Cedex,France

Recei V ed:June29,2004;In Final Form:September2,2004

Discrete nickel catalyst nanoparticles(NP)have been used to grow isolated single-walled carbon nanotubes

by chemical vapor deposition(CVD).Atomic force microscopy(AFM),transmission electron microscopy

(TEM)and micro-Raman spectroscopy are used to characterize these as-grown,isolated https://www.360docs.net/doc/7815576215.html,ing a

conventional setup,we are able to produce narrow diameter distribution single-walled carbon nanotubes

(SWNTs)(1.45(0.5nm),showing a clear mismatch with the larger broad size distribution of the catalyst

nanoparticles(4.7(1.5nm).Investigation of the growth conditions allowed us to synthesize,starting with

the same catalyst materials,various carbon nanotube(CNT)samples with a wide range of characteristics.

The advantages of such a versatile CVD method for the study of physical properties at the single nanotube

level are illustrated by means of two prospective studies on SWNTs mechanical and vibrational properties.

Introduction

Carbon nanotubes(CNTs)have attracted a great deal of interest in the field of fundamental physics at the single molecule level and because of their tremendous potential for future technological applications.1Because the properties of CNTs in general strongly depend on their structural characteristics(e.g., diameter,chirality,length),the control and characterization of these parameters is of major interest when their physical properties are studied.2Among the production methods,chemi-cal vapor deposition(CVD)offers certain advantages in terms of large-scale production,low reaction temperature,and control-lable fabrication.3,4Furthermore,direct CVD growth over supported catalysts allows the production of isolated individual single-walled carbon nanotubes(SWNTs),5,6whereas nanotubes fabricated by arc discharge7,8or laser ablation9,10are produced as bundles.On the other hand,there have been several reports indicating that controlling the size of the catalyst particle allows the control of CNT diameter.11-19However,whether the size of the catalyst particle is a key factor for SWNTs growth is still debated.20-26

One of the mostly used strategies for CVD growth of nano-tubes consists of the preparation of catalyst particles by“wet”processes and in their deposition onto a substrate.In contrast to iron or iron-alloy particles,no report has been dealing with the use of chemically synthesized nickel nanoparticles as SWNT precursors in a supported-catalyst CVD process.5,17,27-29It can be important to diversify CNT precursors,as some of them(like iron)are notorious contaminants to silicon-based devices.

Here,we demonstrate the ability to grow isolated single-walled carbon nanotubes with a relative narrow diameter distribution from discrete nickel nanoparticles with a relatively broad size distribution by catalyst-supported thermal CVD.By varying the synthesis parameters we are able to produce a wide range of nanotube samples with controlled characteristics:small or large diameter SWNTs,with a narrow or broad diameter distribution,a mix of SWNTs and multiwalled carbon nanotubes (MWNTs)and even ultralong SWNTs.Also,the role of the catalyst nanoparticles in nanotube growth will be discussed. Finally,we illustrate the potentialities of such a versatile nanotube synthesis with two short pre-studies of mechanical and vibrational properties at the single nanotube level. Experimental Section

Catalyst Preparation.Nickel nanoparticles were prepared from organometallic precursors as described elsewhere.30Briefly speaking,Ni(COD)2was decomposed in CH2Cl2under H2 atmosphere in the presence of poly(vinylpyrrolidone)(PVP). Evaporation of the solvent led to a black shiny powder of zerovalent nickel colloid.Ni nanoparticles were redispersed in ethanol and deposited on silicon substrates with a100nm thermally grown oxide layer by spin coating(45Hz for1min). Synthesis of Carbon Nanotubes.CVD growth of carbon nanotubes was carried out using ethylene or methane as carbon feedstocks.The supported catalysts were placed at the center of a3′′quartz tube reactor and heated in air up to650°C(at a rate of20°C/min),and then under a flow of argon until the synthesis temperature was reached(750-850-950°C for ethylene and900°C for methane).Depending on the carbon precursor,the following gas mixtures were introduced for1h:

*To whom correspondence should be addressed.E-mail:paillet@

gdpc.univ-montp2.fr.

17112J.Phys.Chem.B2004,108,17112-17118

10.1021/jp0471711CCC:$27.50?2004American Chemical Society

Published on Web10/12/2004

(1)12sccm C2H4/800sccm H2/1200sccm Ar,(2)600sccm CH4/600sccm H2.Finally,the reactive species were purged by using a flow of Ar and the reactor was cooled to room temperature.

Characterization.Ni nanoparticles and carbon nanotubes were characterized by atomic force microscopy(AFM)(Dimen-sion3100,Nanoscope IIIa,Digital Instruments)operating in the tapping mode with commercial Si cantilevers(spring constant～35N/m)and integrated tips.

A recently developed technique31was used to observe nanotubes,which were grown on the substrate,by high resolution transmission electron microscopy(HR-TEM)(Philips CM200,100keV).Shortly,a metallic structure is created by standard lithography methods on the surface.Next,the substrate is cleaved close to this structure.Then,it is etched in such a way that a part of the structure is reaching out over the edge of the substrate.This free-standing structure allows TEM imaging of freely suspended carbon nanotubes sticking to the metal pads. Raman spectra were collected under ambient conditions in a backscattering geometry on a triple subtractive Jobin-Yvon T64000spectrometer(1800g?mm-1,monochromator and spectrograph stages,i.e.,resolution of2cm-1),equipped with an optical microscope(×100objective,～1μm diameter spot on the sample)and a liquid nitrogen cooled CCD detector.Two excitation lines of the Ar+Kr+laser have been used:514.5nm (2.41eV)and647.1nm(1.92eV).The positioning of the laser spot on the sample was monitored with a piezoelectric nano-positioner.

Results and Discussion

Figure1a shows a typical AFM image of a Si substrate with ～100Ni NP/μm2after spin-coating deposition.The original clean Si surface was very smooth with a root-mean-square surface roughness of～0.15nm.The corresponding nanoparticle size distribution before synthesis(Figure3a)indicates a relatively broad size disparity centered at4.7nm with a standard deviation(SD)of1.4nm.The nanoparticle density was tuned

by adjustment of the ethanol stock solution concentration.

Figure1b shows a representative AFM image of nanotubes

grown at850°C from Ni nanoparticles(100NP/μm2)on a

silicon substrate with a100nm thermally grown oxide layer

using ethylene as carbon precursor.Most of the nanotubes have

lengths between500nm and a couple of micrometers,whereas

the shortest measure only tens of nanometers and the longest ～15-20μm(see Figure1b).All nanotubes had a catalyst nanoparticle attached to one extremity.It was proposed11that

such a wide range of lengths could result from catalyst poisoning

effects,disparity among the particles in size,and local variations

of particle-support interactions.

The nanotubes grown from dispersed catalyst nanoparticles

are mostly isolated and only forming small bundles of two or

three tubes when their local density on the substrate is high

enough to enable two nanotubes coming from different nano-

particles to bind together via van der Waals force(vdW).The

growth of two nanotubes originating from the same nanoparticle

is rarely observed.

As revealed by HRTEM(Figure2a,b),the nanotubes are

single-walled with a well-defined wall structure and isolated.

Figure2a shows a high-magnification TEM image of an

individual1.4nm nanotube,partially covered with amorphous

carbon.Self-decomposition of ethylene at high temperature is

likely to be at the origin of this coating.32As previously reported

by other groups,11,20,21only a small portion of the nanoparticles

generates nanotubes.The synthesis yield can be estimated by

the ratio of the number of grown nanotubes on the total number

of nanoparticles present in a given area.Depending on the

experiments,it varies from<0.5%to4%.HRTEM investiga-

tions revealed that inactive nanoparticles during the nanotubes

synthesis were covered by an amorphous carbon layer(see

Figure2a).Such poisoning may explain the low ratio of active

nanoparticles.

The diameters of the nanotubes grown at850°C were

estimated from their heights on AFM topographic images(see

Figure3e).The distribution was found to be quasi Gaussian

with only a couple of nanotubes with diameters larger than3

nm.The average diameter is1.4nm with a SD inferior to0.4

Figure1.(a)AFM topography image of discrete Ni nanoparticles (bright dots,～100NP/μm2)deposited by spin-coating on silicon substrate.(b),(c)AFM topography images of nanotubes grown from Ni nanoparticles on silicon substrates with a100nm thermally grown oxide layer using ethylene:(b)from～100NP/μm2at850°C,(c)from ～50NP/μm2at850°C,(d)from～100NP/μm2at950°C.Figure2.(a)and(b)TEM images of as-grown SWNTs using ethylene at850°C.Scale bars are5nm.

Versatile Synthesis of SWNTs from Ni Nanoparticles J.Phys.Chem.B,Vol.108,No.44,2004

17113

nm (i.e.,>80%of the nanotube diameters lied between 1and 2nm).AFM was preferred to HRTEM for the diameter distri-bution measurement due to the higher amount of data accessible by AFM.However,the AFM data were confirmed by HRTEM diameter measurements of ～30nanotubes (not shown here).It is important to note that all nanotubes (even those with the largest diameters)are found to be single-walled,as illustrated by the HRTEM image of a 3.2nm SWNT (Figure 2b).

As a result,we found a heavy mismatch between the nanoparticles and nanotubes size distribution.Two cases have been reported in the literature.In the first,the catalyst size is shown to be one of the critical factors that allows control of nanotube diameter.11-19In the second,some authors pointed out discordance between nanoparticles and nanotubes sizes,20-26,33and some of them assumed that the carbon supply is the key factor.By lowering the carbon supply,the reaction can be changed from a diffusion-limited to a supply-limited one,which leads to a preferential MWNT or SWNT production,respec-tively.12,34However,and contrary to the present study,SWNTs were in these former works found bundled,suggesting that many small SWNTs could originate from the same bigger nanoparticle (showing similarities with other synthesis techniques such as arc discharge or laser ablation).

To precisely understand the relationship between the nanotube diameter and the size of its generating nanoparticle,it would be necessary to image both simultaneously.11However,the approach we used for imaging the nanotubes by HRTEM does not presently allow us to depict the ends of a given nanotube.

So,we first compare,by AFM,an area pre-located thanks to lithographic markers before and after CVD at 850°C (Figure 4a,b).Figure 4b shows that most of the nanoparticles did not change their positions or their heights.Especially,the nano-particle attached to the SWNT is static on the surface whereas the nanotube has grown away by the base-growth mecha-nism.11,20The stability of nanoparticle size was confirmed by statistical analysis of their heights after synthesis at 850°C (Figure 3b).By AFM measurement (not shown here),a global tendency for the nanotube diameter increase with the attached nanoparticle size has been found.However,we also observed that a 6nm nanoparticle could generate a 1nm nanotube as well as a 3nm one.This result implies that the nanoparticle size is only an upper limit for the generated nanotube diameter.

Figure 3.(a)-(c)Histograms of ～2000Ni nanoparticle diameters measured by AFM as deposited (a),after CVD at 850°C (b),and at 950°C (c).(d)-(f)Histograms of ～200nanotube diameters measured by AFM after CVD synthesis using ethylene at 750°C (statistic on only 50tubes in this case)(d),850°C (e),and 950°C (f).The solid lines correspond to Gaussian fits.

Figure 4.AFM topography images of the same area before (a)and after CVD using ethylene at 850°C for 1h (b),confirming the base-growth of SWNT on the surface.Arrows are guide for the eyes,pointing the same nanoparticle (3.6and 3.5nm respectively)on both images.The generated nanotube is 1.6nm in diameter.

17114J.Phys.Chem.B,Vol.108,No.44,2004Paillet et

al.

This upper limit can be further lowered to～3-4nm regarding to the comparison of the nanotube and the capsule energy displayed by Hafner et al.34The lower limit of nanotube diameter can be estimated from the critical radius of the nucleus (r*).From eq5in ref36,and assuming that the Ni nanoparticles are in a liquid state,r*～0.95-1.25nm for a temperature of 850°C and saturation coefficients of the metal-carbon solution of4and 1.5,respectively.These limits are in reasonable agreement with our experimental results.According to these assumptions,(1)the nanotube diameter fits the nanoparticle size when it is decreased close to the value of r*11,15,16and(2)the synthesis parameters have to be adjusted to get an agreement between both sizes when the nanoparticle size increases.12,18 Influence of the Temperature.We have studied the effect of synthesis temperature on the nanotube produced.Figure1d shows a typical view of nanotubes grown from Ni nanoparticles (100NP/μm2)at950°C.In contrast to850°C synthesis where nanoparticle density and distribution were left unchanged(Figure 3a,b),the nanoparticle density at950°C was divided by a factor of～2,their height distribution was broadened(SD2.25nm) and shifted(mean diameter)5.3nm),and its shape was log-normal(Figure3c)after the CVD process.Likewise,the nanotube diameter distribution(Figure3f)is more than twice as broad(SD～0.85)and shifted(average diameter～1.8nm). Its shape is log-normal with almost45%of the nanotube diameters above2nm and more than10%above3nm.The melting temperature of metal particles smaller than100nm is known to be significantly lower than the bulk melting temper-ature.According to this,5nm nickel particles are predicted to melt at900°C.35At950°C,surface diffusion and collusion of nanoparticles are promoted.As a consequence,their aggregation occurs and an increase of the particle size is observed.

Comparing the results obtained at850and950°C,although the nanotubes in both cases do not mimic the nanoparticles size distribution,the same trend is followed by nanoparticle and nanotube diameter distributions.These results indicate a more complex relation between the catalyst nanoparticles and the SWNTs size distribution.Further studies are needed to fully clarify the situation.

For lower temperature(i.e.,750°C),we observed larger tubes with a mean diameter of3nm with a SD of1.1nm,as shown on Figure3d.This reproduces quite well the result obtained with iron nanoparticles by C.L.Cheung et al.12in close work conditions.Because there are mostly no changes in the particle size up to850°C,the drastic changes of the nanotube diameter between750and850°C are attributed to thermodynamic effects on ethylene decomposition,metal nanoparticle properties,and nanotube nucleation.For nanotubes,the critical radius of carbon nucleus decreases with increasing temperature,36i.e.,smaller radii(higher curvature)require more energy(higher temperature) to be formed.However,no simple considerations can explain the lack of large diameter tubes observed for the850°C growth. We believe that the nature of the metal used(involving melting temperature,carbon solubility,carbon diffusion coefficient,...) may play a role here.35

Carbon Feedstock.Preliminary studies were carried out with various C2H4flows(from1to300sccm)(not shown here). First,the mean nanotube diameter is found to decrease with increasing C2H4partial pressure.These results are in good agreement with the decrease of critical radius of the nucleus with the increase of the saturation coefficient of the metal-carbon solution.With the largest flows,thicker MWNTs are mostly produced.This indicates that the reaction turns from supply-limited to diffusion-limited.These preliminary results will be discussed in detail elsewhere.To study the effect of the carbon precursor,we have investigated growth using methane instead of ethylene.Methane is more energetically stable than ethylene and should produce less amorphous carbon than ethylene at the same synthesis temperature.32Using the same growth duration(1h),the synthesis yield is found to be higher than with ethylene(～10%)and the nanotubes produced are much longer.Among them,a significant fraction is observed to be very long,with lengths>100μm.Figure5a is a zoom-in on an oval-shaped loop(diameter～2μm)characteristic of such ultralong nanotubes.37The particle size distribution after methane-based synthesis at900°C is comparable to the one with ethylene at950°C(Figure3c).On the other hand,the nanotube diameters are totally different,as illustrated in Figure 5b.Contrary to ethylene synthesis,the one using methane is not selective:both small SWNTs(down to～1nm)and MWNTs(>10nm)are produced.These results show that,under these synthesis conditions,the limiting step is(1)the carbon supply for the smaller nanoparticles that are in a liquid state at 900°C(i.e.,with diameters lower than～7nm for Ni-C at the eutectic concentration)and(2)the diffusion of carbon for the bigger ones that are in a solid state at900°C(i.e.,with diameters higher than～7nm).

Substrates.We also carried out experiments with other substrate materials,namely pure silica and silicon wafers.No drastic changes in nanotube characteristics were noted when pure silica substrates were used instead of thermally oxidized silicon(Figure6a).The synthesis yield(～10%)is enhanced probably due to the lower surface roughness.On the contrary, the nanotube density is dramatically decreased when using pure silicon with only a native oxide layer(Figure6b).We observed that the initial density of particles is divided by more than10 Figure5.AFM topography images of nanotubes grown from Ni nanoparticles on silicon substrates with a100nm thermally grown oxide layer using methane at900°C.(a)shows an oval-shaped loop characteristic of an ultralong nanotube(L>100μm).(b)shows two extremes of the nanotube diameter distribution:a10nm MWNT on the left and a1.3nm SWNT on the right,originated from15and6.5 nm nanoparticles,respectively.

Figure6.AFM topography images of nanotubes grown from Ni nanoparticles using ethylene at850°C on different substrates:(a)pure silica with～50NP/μm2and(b)pure silicon with～200-300NP/μm2 before synthesis and～30NP/μm2after.

Versatile Synthesis of SWNTs from Ni Nanoparticles J.Phys.Chem.B,Vol.108,No.44,2004

17115

times after the CVD.Increasing the density of deposited nanoparticles can lead to an appreciable density of nanotubes,but their average length is still lower than on silica.We attribute this to interdiffusion and the formation of defined binary compounds of Ni x Si y that removes most of the particles from the surface.24,38,39It is worth noting that,at a given temperature,the nanotube diameter distribution is independent of the substrate used.

Nanoparticle Density.Finally,we tested various nanoparticle densities:40,100,and up to several hundred NP/μm 2.Figure 1c depicts typical results with 40NP/μm 2where a density of 0.1-0.2SWNTs/μm 2is obtained,i.e.,a yield of 0.4%.With 100NP/μm 2,we obtained ～3SWNTs/μm 2,i.e.,a yield of 3%(see Figure 1b).If the nanoparticle density is too high (several hundreds),the samples turned black after CVD process due to deposition of amorphous carbon and big filamentous structures.This could be understood in terms of nanoparticle aggregation.Even if the increase of synthesis yield with the nanoparticle density (from 30-50to 100-200NP/μm 2)was found to be reproducible,the reason of this remains unclear.

Summary.We obtained,in a controlled manner,several kinds of nanotube samples by adjusting the synthesis conditions.Such a variety of sample characteristics are interesting for the study of physical properties of individual nanotubes.Relatively dense samples are of particular interest for studying individual nanotubes networking,whereas low-density samples are very well adapted to study the properties of isolated individual SWNTs.We hereafter discuss an example of each case.

Mechanical Information.On dense samples,SWNTs are likely passing one on the other and crosses can be observed on as-grown SWNTs with no AFM manipulation.It has been demonstrated that mechanical information can be extracted from the behavior of nanotubes that encounter a topographic obstacle on a surface.40-42We are focusing here on the AFM investiga-tion of the strain of crossing SWNTs.In this configuration,an elastic compression of both tubes at their contact results from the attractive forces between the upper nanotube and the surface and from its flexion around the lower one.Figure 7a shows details of the intersection point of two ～3.2nm SWNTs (see profiles 7b and 7c).At the crossing point the height measured is 4.1nm (see profile 7d),well below the sum of nanotube diameters (～6.4nm).This leads to 36%compression for both nanotubes.Because the interpenetration of the tube is higher than their radius,we can estimate the contact area to their cross section:A )πR 2～8nm 2.The in-plane shear modulus for a nanotube of 3.2nm was calculated to be ～1.1GPa.43This leads to a stress:σ)2 0G 23～0.8GPa or an estimated force F of 6nN (F )σA ).This result is in relative agreement with the order of magnitude predicted by molecular mechanics simula-tions.42A more exhaustive study over a large number of crosses with a wide range of nanotube diameters is in progress in our group.

Raman Spectroscopy.Our versatile CVD synthesis appears suitable for the study of the vibrational properties at the individual nanotube level.

Resonant Raman spectroscopy has provided a powerful tool for the observation of diameter-selective phonon modes of single-walled carbon nanotube (SWNTs).44,45At the single resonance level,the unique electronic properties of SWNTs allow selective Raman studies of metallic and semiconducting nanotubes as a function of laser excitation energy.46The Raman spectra of SWNTs are dominated by radial breathing modes (RBM)below 300cm -1and by tangential modes (TM)in the high-frequency region (1400-1700cm -1).The Raman shift of RBM (νRBM in cm -1)for an individual SWNT is related to its diameter (d in nm)by νRBM )248/d.47A typical spectrum of an isolated nanotube is shown in Figure 8(bottom curves).The laser energy used (1.92eV)and the position of the RBM (νRBM )187cm -1)indicate that this nanotube is metallic with a diameter around 1.33nm.This isolated individual SWNTs comes from a low-density sample (<1nanotube/μm 2)synthe-sized by the CVD process described here.The nanotube was pre-located by AFM using lithographic markers visible with the micro-Raman microscope.We are presenting for comparison the Raman responses (RBM and TM)of a pristine SWNTs’bundles sample (Figure 8,top curves)synthesized by the arc-discharge method 48at the same laser excitation wavelength.The Raman response of SWNTs is known to be strongly dependent from their structural characteristics (diameter and chirality)and from their orientation relative to the laser polarization.In addition,the van der Waals interactions between nanotubes in bundles can also cause drastic changes in both RBM and TM

Figure 7.(a)AFM topography image of two as-grown crossing nanotubes.(b)and (c)Cross sections of each nanotubes:both ～3.2nm in diameter.(d)Profile at the crossing point,height of ～4.1nm,showing a 2.3nm flattening of the nanotubes at their intersection.

17116J.Phys.Chem.B,Vol.108,No.44,2004Paillet et

al.

positions and profiles.In this context,it is obvious that the assignment of the Raman modes and its correlations with a given nanotube structure require a study at the single nanotube level.49As an illustration,we only want to point out the unambiguous dependence of the TM shape on whether the metallic SWNTs are in bundles or individuals.TM Raman scattering from metallic SWNT bundles shows a low-frequency broad asym-metric component (around 1550cm -1)whose profile is well-described by a Breit -Wigner -Fano (BWF)line shape.44-46This component vanishes when the metallic nanotube measured is isolated.This result highlights that the BWF profile is extremely enhanced in bundles and almost nonexistent in isolated SWNTs,as predicted theoretically 50,51and previously reported.51-53Conclusions

We have grown isolated single-walled carbon nanotubes from discrete catalyst nickel nanoparticles by CVD.We tested several synthesis parameters influencing nanotube growth:nanoparticle size distribution and density,temperature,carbon supply,and substrate.We demonstrated that it is possible to obtain,in a controllable way,from the same relatively broad distribution of nickel nanoparticles (4.7(1.5nm),(1)a relatively narrow distribution of small diameter SWNTs (1.45(0.5nm),(2)a relatively broad distribution of larger SWNTs (1.8(0.9nm),and (3)a mix of SWNTs (from 1nm up to 6nm)and MWNTs (up to >10nm).We put forward that a clear mismatch between nanoparticle and nanotube diameters does exist when a suf-ficiently low carbon supply is used.In our case,the nanoparticle size appears only as an upper limit for the generated nanotube diameter.The lower limit of nanotube diameter is estimated from the critical radius of the nucleus (i.e.,between 0.95and 1.25nm at 850°C).In summary,the thermodynamical parameters (saturation coefficient of the metal -carbon solution,temperature,...)are believed to mainly control the SWNT diameter.Further understanding of the nanoparticle/nanotube relationships in this situation will be important to fully under-stand and control the nanotube nucleation and growth processes.In a second part,we illustrated the large possibilities of such a versatile CVD method for the study of physical properties at the single nanotube level with two prospective examples.First,the analysis of the crossing behavior of two 3.2nm SWNTs revealed almost 40%of compression at the intersection.Second,a simplified comparison of a SWNT pristine Raman spectrum with the one of an individual metallic SWNT supports the vanishing of the BWF component with the decrease of the bundle size.More complete studies are in progress in our group.Acknowledgment.We thank Dr.F.Marchi,Dr.P.Bernier’s group,https://www.360docs.net/doc/7815576215.html,urent’s group,and Dr.M.Ramonda for their

technical supports and helpful discussions,and Dr.X.Devaux,Dr.G.Saussines,and Dr.C.S.Lee for helpful discussions.This research was supported by the NOI French CNRS Program.References and Notes

(1)Baughman,R.H.;Zakhidov,A.A.;de Heer,W.

A.Science 2002,297,787-792.

(2)Dai,H.J.Acc.Chem.Res.2002,35,1035-1044and references therein.

(3)Dai,H.;Kong,J.;Zhou,C.;Franklin,N.;Tombler,T.;Cassell,A.;Fan,S.;Chapline,M.J.Phys.Chem.B 1999,103,11246-11255.

(4)Dai,H.J.Surf.Sci.2002,500,218-241and references therein.(5)Kong,J.;Cassel,A.M.;Dai,H.Chem.Phys.Lett.1998,292,567-574.

(6)Kong,J.;Soh,H.T.;Cassel,A.M.;Quate,C.F.;Dai,H.Nature 1998,395,878-881.

(7)Iijima,S.;Ichihashi,T.Nature 1993,363,603-605.

(8)Bethune,D.S.;Kiang,C.H.;de Vries,M.S.;Gorman,G.;Savoy,R.;Vazquez,J.;Beyers,R.Nature 1993,363,605-607.

(9)Guo,T.;Nikolaev,P.;Thess,A.;Colbert,D.T.;Smalley,R.E.Chem.Phys.Lett.1995,243,49-54.

(10)Thess,A.;Lee,R.;Nikolaev,P.;Dai,H.;Petit,P.;Robert,J.;Xu,C.;Lee,Y.H.;Kim,S.G.;Rinzler,A.G.;Colbert,D.T.;Scuseria,G.E.;Tomanek,D.;Fischer,J.E.;Smalley,R.E.Science 1996,273,483-487.(11)Li,Y.;Kim,W.;Zhang,Y.;Rolandi,M.;Wang,D.;Dai,H.J.Phys.Chem.B 2001,105,11424-11431.

(12)Cheung,C.L.;Kurtz,A.;Park,H.;Lieber,C.M.J.Phys.Chem.B 2002,106,2429-2433.

(13)Siegal,M.P.;Overmyer,D.L.;Provencio,P.P.Appl.Phys.Lett.2002,80,2171-2173.

(14)Bonard,J.-M.;Chauvin,P.;Klinke,C.Nano Lett.2002,2,665-667.

(15)An,L.;Owens,J.M.;McNeil,L.E.;Liu,J.J.Am.Chem.Soc.2002,124,13688-13689.

(16)Choi,H.C.;Kim,W.;Wang,D.;Dai,H.J.Phys.Chem.B 2002,106,12361-12365.

(17)Sato,S.;Kawabata,A.;Nihei,M.;Awano,Y.Chem.Phys.Lett.2003,382,361-366.

(18)Fu,Q.;Huang,S.;Liu,J.J.Phys.Chem.B 2004,108,6124-6129.

(19)Han,S.;Yu,T.;Park,J.;Koo,B.;Joo,J.;Hyeon,T.;Hong,S.;Im,J.J.Phys.Chem.B 2004,108,8091-8095.

(20)Su,M.;Li,Y.;Maynor,B.;Buldum,A.;Lu,J.P.;Liu,J.J.Phys.Chem.B 2000,104,6505-6508.

(21)Li,Y.;Liu,J.;Wang,Y.;Wang,Z.L.Chem.Mater.2001,13,1008-1014.

(22)Ago,H.;Ohshima,S.;Uchida,K.;Yumura,M.J.Phys.Chem.B 2001,105,10453-10456.

(23)Yoon,J.Y.;Bae,J.C.;Baik,K.H.;Cho,S.;Lee,S.-J.;Song,M.K.;Myung,S.N.Chem.Phys.Lett.2002,366,109-114.

(24)Jung,Y.J.;Homma,Y.;Ogino,T.;Kobayashi,Y.;Takagi,D.;Wei,B.;Vajtai,R.;Ajayan,P.M.J.Phys.Chem.B 2003,107,6859-6864.

(25)Nishide,D.;Kataura,H.;Suzuki,S.;Okubo,S.;Achiba,Y.Chem.Phys.Lett.2004,392,309-313.

(26)Kohno,M.;Orii,T.;Hirasawa,M.;Seto,T.;Murakami,Y.;Chiashi,S.;Miyauchi,Y.;Maruyama,S.To be published in Appl.Phys.A .(27)Colomer,J.-F.;Bister,G.;Willems,I.;Ko ′nya,Z.;Fonseca,A.;Tendeloo,G.V.;Nagy,https://www.360docs.net/doc/7815576215.html,mun.1999,14,1343-1344.(28)Lee,J.H.;Heo,J.N.;Yi,W.K.;Jeong,T.W.;Yu,S.;Lee,C.S.;Kim,W.S.;Lee,T.Y.;Yoo,J.B.;Kim,J.M.Physica B 2002,323,321-323.

(29)Seidel,R.;Duesberg,G.S.;Unger,E.;Graham,A.P.;Liebau,M.;Kreupl,F.J.Phys.Chem.B 2004,108,1888-1893.

(30)Ould Ely,T.;Amiens,C.;Chaudret,B.;Snoeck,E.;Verelst,M.;Respaud,M.;Broto,J.-M.Chem.Mater.1999,11,526-529.

(31)Meyer,J.C.;Obergfell,D.;Yang,S.;Yang,S.F.;Roth S.Appl.Phys.Lett.,in press.

(32)Tibbetts,G.G.Academic:Amsterdam,1990.

(33)Huang,S.;Woodson,M.;Smalley,R.E.;Liu,J.Nano Lett.2004,4,1025-1028.

(34)Hafner,J.H.;Bronikowski,M.J.;Azamian,B.R.;Nikolaev,P.;Rinzler,A.G.;Colbert,D.T.;Smith,K.A.;Smalley,R.E.Chem.Phys.Lett.1998,296,195-202.

(35)Moisala,A.;Nasibulin,A.G.;Kauppinen,E.I.J.Phys.:Condens.Matter 2003,S3011-S3035.

(36)Kuznetsov,V.L.;Usoltseva,A.N.;Chuvilin,A.L.;Obraztsova,E.D.;Bonard,J.-M.Phys.Re V .B 2001,64,235401-235407.

(37)Kim,W.;Choi,H.C.;Li,Y.;Wang,D.;Dai,H.Nano Lett.2002,2,703-708.

Figure 8.RBM (left)and TM (right)regions of the Raman spectrum excited at 647.1nm (1.92eV)measured on an isolated individual SWNT (bottom curves)and on a pristine SWNT (organized in bundles)sample synthesized by electric-arc method (top curves).

Versatile Synthesis of SWNTs from Ni Nanoparticles J.Phys.Chem.B,Vol.108,No.44,200417117

(38)Jung,Y.J.;Wei,B.Q.;Vajtai,R.;Ajayan,P.M.;Homma,Y.; Prabhakaran,K.;Ogino,T.Nano Lett.2003,3,561-564.

(39)Homma,Y.;Kobayashi,Y.;Ogino,T.;Takagi,D.;Ito,R.;Jung, Y.J.;Ajayan,P.M.J.Phys.Chem.B2003,107,12161-12164.

(40)Biro′,L.P.;Lazarescu,S.;Lambin,P.;Thiry,P.A.;Fonseca,A.; Nagy,J.B.;Lucas,A.A.Phys.Re V.B1997,56,12490-12498.

(41)Hertel,T.;Martel,R.;Avouris,P.J.Phys.Chem.B1998,102, 910-915.

(42)Hertel,T.;Walkup,R.E.;Avouris,P.Phys.Re V.B1998,58, 12870-13873.

(43)Shen,L.;Li,J.Phys.Re V.B2004,69,045414-045410.

(44)Dresselhaus,M.S.;Eklund,P.C.Ad V.Phys2000,49,705-814.

(45)Sauvajol,J.-L.;Anglaret,E.;Rols,S.;Alvarez,L.Carbon2002, 40,1697-1714.

(46)Kataura,H.;Kumazawa,Y.;Maniwa,Y.;Umezu,I.;Suzuki,S.; Ohtsuka,Y.;Achiba,Y.Synth.Met.1999,103,2555-2558.

(47)Jorio,A.;Saito,R.;Hafner,J.H.;Lieber,C.M.;Hunter,M.; McClure,T.;Dresselhaus,G.;Dresselhaus,M.S.Phys.Re V.Lett.2001, 86,1118-1121.

(48)Journet,C.;Maser,W.K.;Bernier,P.;Loiseau,A.;de la Chapelle, M.L.;Lefrant,S.;Deniard,P.;Lee,R.;Fischer,J.E.Nature1997,388, 756-758.

(49)Dresselhaus,M.S.;Dresselhaus,G.;Jorio,A.;Souza Filho,A.

G.;Saito,R.Carbon2002,40,2043-2061.

(50)Kempa,K.Phys.Re V.B2002,66,195406.

(51)Jiang,C.;Kempa,K.;Zhao,J.;Schlecht,U.;Kolb,U.;Basche,T.; Burghard,M.;Mews,A.Phys.Re V.B2002,66,161404.

(52)Kataura,H.;Kumazawa,Y.;Kojima,N.;Maniwa,Y.;Umezu,I.; Masubuchi,S.;Kazama,S.;Ohtsuka,Y.;Suzuki,S.;Achiba,Y.Mol.Cryst. Liq.Cryst.2000,340,757-762.

(53)Bendiab,N.;Almairac,R.;Paillet,M.;Sauvajol,J.-L.Chem.Phys. Lett.2003,372,210-215.

17118J.Phys.Chem.B,Vol.108,No.44,2004Paillet et al.