Spectroscopic and Scanning Probe Studies of a Nondestructive Purification Method forSWNT Suspensions

Jihye Shim,† Pornnipa Vichchulada, Qinghui Zhang, and Marcus D. Lay*Department of Chemistry and Nanoscale Science and Engineering Center (NanoSEC),UniVersity of Georgia, Athens, Georgia 30602

ReceiVed: September 8, 2009; ReVised Manuscript ReceiVed: NoVember 16, 2009

A mild method for the bulk enrichment of high-aspect-ratio single-walled carbon nanotubes (SWNTs), whileavoiding the damaging effects of acid purification methods, has been developed in this group. Various analyticaltechniques (atomic force microscopy, Raman microscopy, near IR, and UV-vis spectroscopy) have beenused in order to attain a better understanding of the correlation between low-G centrifugation cycles and thequality of the SWNT suspension. Atomic force microscopy data of low-density SWNT networks formedfrom these suspensions indicate that the size and occurrence of globular impurities is lessened, while smallbundles of high-aspect-ratio SWNTs remain. Raman spectroscopy is used to verify that this method producessuspensions of purified SWNTs without significantly damaging them. Near-infrared and UV-vis spectroscopyof these suspensions show prominent semiconducting and metallic interband transitions in the SWNTs andremoval of impurities.

Introduction

Because of their unique physical, chemical, and mechanicalproperties,1,2 single-walled carbon nanotubes (SWNTs) haveapplications in a wide variety of scientific disciplines, includingfield emission devices,3 electronic devices,4 chemical sensors,5

batteries,6 hydrogen storage,7 and polymeric composites.7,8

However, most commercial synthesis methods for SWNTgrowth produce a large percentage (up to 60%) of carbonaceousimpurities, such as amorphous carbons, fullerenes, nanocrys-talline graphite, and transition-metal nanoparticles that are usedas a catalyst during the synthesis.9 These impurities have adeleterious effect on the performance of various device struc-tures, such as resistors, transistors, and sensors.10 Therefore, thereis a great need for methods for purification of SWNT soot thatdo not damage the enhanced electrical properties of the SWNTs.

Reported methods for purification, such as gas- or liquid-phase oxidation, significantly damage pristine as-produced (AP)SWNTs. For example, gas-phase oxidation, used for the removalof carbonaceous impurities, usually occurs in an atmospherecontaining a mixture of some of the following gases at elevatedtemperatures: air, oxygen, ozone, hydrogen chloride, chlorine,CCl4, or CO2.11,12 While any single one of these gases was noteffective at purifying SWNT soot, of the dual-gas mixtures used,HCl(g) and H2O(g) were most effective at etching carbonaceousimpurities. However, because the SWNTs are also composedentirely of carbon, these gases also etched the SWNTs, leavingeither hydroxyl or chloride functional groups along the endsand sidewalls. Further, after purification, it is difficult to removesome of these gases from the sidewalls of the SWNTs.Therefore, in addition to creating defects in the SWNTs, theincomplete removal of these gases presents another problem,as electron-donating gases reduce the conductivity of SWNTs,which act as p-type semiconductors.13,14 Additionally, gas-phase

methods have not proven effective for bulk purification ofSWNT soot.

Liquid-phase oxidation processes, in strong acids, such asHNO3, HCl, HClO4, and H2SO4, are an alternative commonlyused for bulk oxidative purification of SWNT soot. Althougheffective at reduction in carbonaceous impurities and metalnanocatalyst particles, acid treatments also attack the sp2

hybridized structure of the SWNTs, thus changing their intrinsicelectronic transport properties.15,16 Therefore, the developmentof nondamaging methods for the purification of SWNTs is ofgreat interest. The most common mild purification methodsinvolve, first, creating an aqueous suspension of SWNTs. Asurfactant is used to increase the solubility of the SWNTswithout the need for chemical modification. Next, SWNTs areseparated from impurities via filtration or centrifugation.17,18

Filtration methods work well for the production of three-dimensional networks of transparent SWNTs. This materialexhibits near metallic conduction, so it has applications similarto indium tin oxide (ITO). For example, when a dilutesuspension of single-walled carbon nanotubes (SWNTs) insurfactant was vacuum-filtered onto a filtration membrane,followed by dissolving the membrane in solvent, the result wasan SWNT film with a thickness of 50-150 nm.19 Although thesefilms were composed of SWNTs that were bundled together,these films exhibited a high transmittance in the visible and near-IR. While this method was effective for formation of SWNTthin films with metallic conduction, the advantage of thepurification and deposition method presented herein is that low-density, two-dimensional networks behave as transparent semi-conductive materials.20,21

Recently, a combination of high-speed centrifugation andmembrane filtration has been used for purification of SWNTspretreated by oxidation with HNO3 or K2S2O8.22,23 The oxidationprocess was used to chemically modify the SWNTs in order toincrease their solubility in H2O. Although bulk separation ofcarbonaceous impurities from SWNTs was accomplished,chemical modification of the SWNTs was crucial to the successof this work.

* To whom correspondence should be addressed. E-mail: mlay@chem.uga.edu.

† Current address: Department of Applied Chemistry, Kyung HeeUniversity, Gyeonggi-do, 446-701, Korea.

J. Phys. Chem. C 2010, 114, 652–657652

10.1021/jp9086738 2010 American Chemical SocietyPublished on Web 12/14/2009

An effective purification method that does not involvechemical modification of SWNTs is ultracentrifugation, whichuses centrifugal forces of at least 100 000 G. This is an effectiveway of removing carbonaceous and metallic impurities fromaqueous suspensions of SWNTs, as impurities usually havehigher densities than short, unbundled SWNTs.24 However, inthis technique, long unbundled SWNTs, which are useful fornumerous electronic applications, are also removed due to theirhigher mass; the average length of SWNTs observed was only200 nm. Therefore, it is important to develop an effectivemethod for purification that removes the various impurities butleaves unbundled, high-aspect-ratio SWNTs in suspension.

This manuscript reports the use of several analytical methodsto study the properties of SWNTs that were purified by multiplecycles of centrifugation at low centrifugal force. This workshows that this purification method is an effective way to obtainbulk samples of purified SWNTs without removal of unbundled,high-aspect-ratio SWNTs; the average length observed by AFMexceeds 1 µm. Raman microscopy and near-IR transmissionstudies of SWNT suspensions confirmed that impurities wereremoved by this method. SWNTs deposited onto Si/SiOx

substrates via laminar flow deposition (LDF) demonstrated thatthe suspensions were enriched in individual, high-aspect-ratioSWNTs.24,25 This deposition method prevents the formation ofSWNT bundles during the deposition process. This provides adistinct advantage over other liquid deposition methods, whichresult in formation of large bundles of SWNTs,22 and allowsdetermination of the effectiveness of the purification methodpresented herein.

Experimental Methods

Preparation and Purification of SWNT Solutions. As-produced AP-grade CarboLex Inc. arc-discharge soot, with aguaranteed purity of 50-70 wt % SWNTs, was used for allexperiments. Solutions of 1 mg/mL of SWNT soot wereprepared in 1% (w/v) sodium dodecyl sulfate (SDS) bydispersion of SWNTs via probe sonication (model 500, Fisher)at 12 W for 30 min, followed by multiple centrifugation(Beckman Microfuge) cycles of 30 min each at 18 000 G. Eachcentrifugation cycle was followed by carefully collecting theupper half of the supernatant. Spectroscopic and scanning probedata were obtained after each centrifugation cycle in order toquantitate the effectiveness of this method for removal ofimpurities and enrichment in high-aspect-ratio SWNTs.

Formation of SWNT Deposits through Laminar FlowDeposition. Si/SiO2 substrates were cut into fragments and thencleaned with compressed CO2. The substrates were thenfunctionalized with a self-assembled monolayer to produce anamine-terminated surface by immersion in a fresh solution of10 mM 3-aminopropyl triethoxysilane in 99.5% ethanol for 45min. Ten deposition cycles were then used to form a low-densitySWNT network. The deposition cycle has been describedelsewhere.26 Briefly, each cycle is defined as depositing theSWNT solution onto the silane-coated Si/SiO2 substrate bydrying in a stream of N2. The substrate was then rinsed withnanopure-H2O and dried under a stream of N2.

Characterization of Purified SWNT Suspensions by AFMand Absorption Spectroscopy. AFM images were obtained inair using intermittent contact mode (Molecular Imaging Pico-Plus). To determine the extent of purification, image analysissoftware (SPIP) was used to create a histogram of the heightmeasurements for each pixel. This is an effective analysismethod because the various impurities were consistently ob-served to have higher height than the average SWNT (∼1.2

nm). For each purification cycle, transmission near-infrared(NIR) spectroscopy (Thermo 6700 Nicolet FT-IR spectroscope)was performed on aqueous suspensions using a quartz cell witha path length of 1 cm. Ultraviolet-visible (UV-vis) absorptionspectroscopy was performed on a Varian Cary 300 spectropho-tometer in a similar quartz cell. Aqueous suspensions, as wellas SWNT networks formed on Si wafer fragments (Si/SiOx),were characterized with Raman spectroscopy. Raman spectrawere recorded on a Renishaw InVia Raman microscope, witha charge-coupled device (CCD) detector, using a diode laserwith an excitation wavelength of 785 nm and 7 mW laser power.For aqueous suspensions, a capillary tube (1.5 × 90 mm) wasused as a sample container. For Raman analysis of two-dimensional networks, SWNTs deposited on Si/SiOx wereexamined without further modification. Analysis of the heightsof bands in the Raman spectra was performed with WiRe 2.0software.

Results and Discussion

Atomic Force Microscopy Studies. Although liquid deposi-tion processes have the major advantage of decoupling theSWNT growth process (which occurs at ∼800 °C) from thedeposition process, a major concern with liquid depositionprocesses is bundle formation during the deposition process.Laminar flow deposition is an effective way to prevent SWNTbundle formation during deposition of SWNT networks.25

Therefore, AFM analysis of deposits formed with LFD allowsquantitation of the change in the density and size of impuritiesas a function of purification cycles, when the number ofdeposition cycles is held constant. The results indicate that anincrease in centrifugation cycles decreases the size and densityof impurities.

Figure 1 shows the effect of centrifugation cycles at low Gon the quality of SWNT networks deposited onto Si/SiOx

substrates. Figure 1a shows a deposit formed from anunpurified SWNT suspension. An average height of 150 nmwas observed, and the surface is largely covered by globularimpurities. This type of deposit is typical of untreatedsuspensions. However, a dramatic improvement in the qualityof deposit was observed after just one centrifugation cycleat low G (Figure 1b). The average height decreased to 55nm due to removal of a significant portion of the larger im-purities. SWNTs are clearly visible in this deposit. As thenumber of centrifugation cycles increases, the average heightobserved continues to decrease. After two centrifugationcycles (Figure 1c), a histogram of the heights in the zdirection showed an average height of 12.5 nm, with asignificant contribution from heights around 10 nm. Ad-ditionally, significant contributions from heights of less than2 nm are present. There is a dramatic shift in the averageheight of the histograms between two and three centrifugationcycles; despite the fact that there is evidence of a bimodaldistribution, for the first time, the average height is below 5nm (Figure 1d). Four centrifugation cycles result in a smallshift of the average of the height histogram, but with a largerrepresentation from heights below 5 nm, as indicated by theasymmetric Gaussian curve observed in Figure 1e. After fivecentrifugation cycles, the average height approaches the rangewithin one may expect to find SWNTs (∼2 nm), with onlya small contribution from slightly larger heights (Figure 1f).

As the average height range for arc-discharge SWNTs is1.2-1.6 nm, this would indicate that, after five centrifugationcycles were performed on the deposition solution, the depositis composed largely of individual SWNTs, with the presence

Nondestructive Purification Method for SWNT Suspensions J. Phys. Chem. C, Vol. 114, No. 1, 2010 653

654 J. Phys. Chem. C, Vol. 114, No. 1, 2010 Shim et al.

of very small diameter globular impurities. This indicatesthat multiple centrifugation cycles are an efficient, nonoxi-dizing way to remove impurities as well as large bundles ofSWNTs.

Raman Spectroscopy of SWNT Deposits and Suspensions.Raman Spectroscopy of SWNT Suspensions. Raman scatteringhas been used for many years as a probe of disorder in thecarbon skeleton of sp2 and sp3 carbon materials.27 There aretwo Raman bands that are typically observed for sp2 carbon;28

the “G band,” which is called a “tangential mode,” is seen inthe range of 1580-1600 cm-1. Its presence indicates a pristine,symmetrical graphene lattice.29 The second band observed isthe “D band,” which is called a “dispersive” band. It is arelatively broad, disorder-induced band in the range of 1300-1370cm-1.28,29 The D band, which is indicative of the presence ofdisordered sp2 hybridized carbon atoms, is caused by C-atomvacancies, impurities in the C lattice, or any other imperfectionsin the graphene lattice.30

Figure 2 shows the Raman spectra of SWNT suspensionstreated with zero (0C, i.e. without centrifugation) to five low-Gcentrifugation cycles (5C). The G and D bands are observed at1591-1595 and 1290-1310 cm-1, respectively. The magnitudeof the IG/ID is an indicator of the purity of SWNTs.31 There isa consistent increase in the IG/ID ratio, as the G band was

observed to increase with each low-G centrifugation cycle (Table1). Indeed, the IG/ID ratio increased from 4.34 to 17.05 over thecourse of five centrifugation cycles. This increase indicatesenrichment in well-ordered sp2 bonds (such as those observedin defect-free SWNTs) and a corresponding reduction in defect-containing carbonaceous impurities.32

Figure 1. AFM micrographs (8 × 8 µm) for deposits formed from solutions at various stages of purification. (a) AFM image and histogram of adeposit formed from an unpurified SWNT suspension with a concentration of 1 mg/mL of SWNT soot. The average height observed on this depositis 150 nm. (b) The average height decreases to 55 nm for a deposit formed from a suspension that underwent one centrifugation cycle. (c) Aftertwo purification cycles, the average height observed is 12.5 nm. The shoulder on the left side of the peak in the histogram is indicative of thepresence of a bimodal distribution, with heights of significantly less than 10 nm strongly represented. (d) After three centrifugation cycles, theaverage height is less than 5 nm, with a significant presence of particles up to 20 nm. (e) After four centrifugation cycles, the average height is 5nm, with the large contribution of the height less than 5 nm. (f) Five low-G centrifugation cycles result in an average height of 2.0 nm, indicatingthe continued removal of various impurities and bundles, without the use of oxidizing treatments. Additionally, the deposit demonstrates enrichmentin high-aspect-ratio SWNTs.

Figure 2. Raman spectra showing the D band (1290-1310 cm-1) andthe G band (1591-1595 cm-1) of the SWNT suspensions beforecentrifugation (0C) and after centrifugation from one to five cycles.(1C-5C). As the number of centrifugation cycles increased, IG/ID ratiosincreased.

Nondestructive Purification Method for SWNT Suspensions J. Phys. Chem. C, Vol. 114, No. 1, 2010 655

Raman Spectroscopy of 2-D SWNT Networks Deposited onSi/SiOx. Confocal Raman microscopy of 2-D SWNT networksdeposited on Si wafer fragments also evidenced a consistentreduction in carbonaceous impurities with increasing low-Gcentrifugation cycles (Figure 3). This trend indicates that thequality of the deposition solution has a direct correlation onthe quality of thin film formed using laminar flow deposition(Table 1); the IG/ID ratio increased from 4.92 to 17.21. Like theSWNT suspensions, the deposits show the highest IG/ID ratioafter five centrifugation cycles. Therefore, Raman spectroscopydata concur with AFM data with regard to the effectiveness oflow-G centrifugation for purification of SWNT soot.

Near-Infrared Spectroscopy of SWNT Suspensions. Theeffectiveness of low-G centrifugation cycles for bulk purificationof SWNT soot was also evaluated by near-infrared (NIR)spectroscopy in transmission mode. NIR spectroscopy has beenreported as an important tool for characterizing the electronicband structure of SWNTs.33 When unbundled SWNTs arepresent in aqueous suspensions, a series of characteristicinterband electronic transitions are observed; semiconductingSWNTs have first and second transitions starting with S11 )2R�/d and S22 ) 4R�/d, whereas the first transitions of themetallic SWNTs appear at M11 ) 6R�/d, where R is the C-Cbond length (0.1424 nm); � ) ∼2.9 eV, the transfer integralbetween π orbitals (an interaction energy between neighboringC atoms); and d is the SWNT’s diameter (nm).33,34 Theseinterband transitions produce prominent features in the NIRspectral range, provided the electronic transitions are notquenched by inter-SWNT interactions that occur in bundles, sothey can be a basis for the evaluation of the enrichment of thesuspension in unbundled SWNTs.

Figure 4 shows NIR spectra of SWNT suspensions at varyingdegrees of processing. The S11, S22, and M11 transitions wereobserved at 1150, 980, and 750 nm, respectively. All of theseabsorption bands are suppressed in the untreated SWNTsuspension. However, the intensities of these bands increase withincreasing low-G centrifugation cycles.

Studies by Haddon et al., have shown that, for acid-treatedSWNT soot, the second semiconducting interband transition,S22, can be used for purity evaluation because it is less affectedby doping during chemical processing.9,34 This method was usedin this study for the S22 transition, which was used to quantifythe extent of enrichment in unbundled SWNTs via the ratio A(S)/A(T), where A(S) is the area of the S22 absorption feature afterbaseline subtraction and A(T) is the total area under the spectralcurve.

In Table 2, it is shown that multiple centrifugation cyclesincrease the purity ratio. The one-cycle centrifugation had apurity of 96.52%. The purity of the SWNT suspensions aftertwo and three cycles of centrifugation increased to 97.22% and97.43%, respectively. Four cycles showed a higher increase inpurity compared with those that were subjected to fewer cycles.In five centrifugation cycles, the purity was remarkably high at99.96%.

UV-vis Absorption Spectroscopy of SWNT Suspensions.The absorbance of SWNT suspensions, depending on centrifu-gation cycles, was monitored by UV-vis absorption spectros-copy. The spectra showed that more centrifugation cyclesdecreased the absorbance of SWNTs. Current synthetic methodsproduce the mixtures of metallic and semiconducting SWNTs.33

In the UV-vis range, metallic SWNT showed first transition,M11, in 700-800 nm and the semiconducting transitions wereshown in the NIR range.33,34

In Figure 5, the change of absorbance at 750 nm was plottedin the SWNT from zero (i.e., without centrifugation) to fivecentrifugation cycles. Figure 5 shows an exponential decay andleveling off with the decrease of the absorbances from 2.576 to0.175, as centrifugation cycles increased. The uncentrifugedSWNT suspension had the highest absorbance due to bundlesand other impurities. This indicates that most of the largercontaminants are removed in the first centrifugation cycle. Thiscorroborates the observations made with AFM, where much ofthe larger amorphous C contaminants are removed after the firstcentrifugation cycles. Subsequent centrifugation cycles thenremove bundles of SWNTs, as well as smaller globularcontaminants.

Conclusions

Multiple cycles of centrifugation at low G are an efficientand nonoxidizing method for the purification of SWNT soot.

Figure 3. Raman spectra, showing the D band (1290-1310 cm-1)and G band (1591-1595 cm-1) for SWNT deposits formed fromsuspensions treated with no centrifugation (0C) up to five (1C-5C).As the number of centrifugation cycles increased, IG/ID ratiosincreased.

TABLE 1: IG/ID Ratios of the SWNTs in RamanSpectroscopy

OC 1C 2C 3C 4C 5C

SWNT suspensions 4.34 10.11 11.29 12.93 14.47 17.05SWNT deposits 4.92 10.23 12.46 13.97 14.54 17.21

Figure 4. NIR spectra showing metallic (M11), second semiconducting(S22), and first semiconducting (S11) transitions at 750-800, 980, and1150 nm, respectively, of the SWNT suspensions before and aftercentrifugation.

TABLE 2: The Ratios, A(S)/A(T), of the SWNT Suspensionsin NIR Spectroscopy

sample OC 1C 2C 3C 4C

A(S)/A(T) × 100% 96.52% 97.22% 97.43% 99.10% 99.96%

656 J. Phys. Chem. C, Vol. 114, No. 1, 2010 Shim et al.

Experimental results showed that most of the heavier/largercontaminants are removed in the first centrifugation cycle,whereas subsequent cycles remove bundles of SWNTs andsmaller globular contaminants. This results in suspensionsenriched in long unbundled/undamaged SWNTs. The largestbundles of SWNTs observed by AFM after five centrifugationcycles were likely composed of about two SWNTs, as the totalheight was 2.6 nm. NIR study showed more prominent metallicand semiconducting transition interbands with a higher purityratio, A(S)/A(T), according to the increasing number of cen-trifugation cycles. Also, both deposits and suspension-phaseRaman spectroscopy provided higher ratios of IG/ID as thenumber of centrifugation cycles increased, indicating a reducedpresence of defect sites due to C vacancies. Therefore, this mildpurification technique can be applied for efficient and nonde-structive purification of SWNT soot.

Acknowledgment. The authors gratefully acknowledge fi-nancial support from the National Science Foundation throughNSF Grant No. DMR-0906564.

References and Notes

(1) Iijima, S. Nature 1991, 354, 56.(2) Snow, E. S.; Perkins, F. K.; Houser, E. J.; Badescu, S. C.; Reinecke,

T. L. Science 2005, 307, 1942.(3) Fan, S. S.; Chapline, M. G.; Franklin, N. R.; Tombler, T. W.;

Cassell, A. M.; Dai, H. J. Science 1999, 283, 512.(4) Collins, P. G.; Zettl, A.; Bando, H.; Thess, A.; Smalley, R. E.

Science 1997, 278, 100.(5) Turyan, I.; Mandler, D. Anal. Chem. 1997, 69, 894.

(6) Che, G. L.; Lakshmi, B. B.; Fisher, E. R.; Martin, C. R. Nature1998, 393, 346.

(7) O’Connell, M. J., Ed.; Carbon Nanotubes: Properties and Applica-tions; CRC Taylor & Francis: Boca Raton, FL, 2006.

(8) Zhang, T.; Mubeen, S.; Bekyarova, E.; Yoo, B. Y.; Haddon, R. C.;Myung, N. V.; Deshusses, M. A. Nanotechnology 2007, 18, 165504.

(9) Nikolaev, P.; Bronikowski, M. J.; Bradley, R. K.; Rohmund, F.;Colbert, D. T.; Smith, K. A.; Smalley, R. E. Chem. Phys. Lett. 1999, 313,91.

(10) Itkis, M. E.; Perea, D. E.; Jung, R.; Niyogi, S.; Haddon, R. C. J. Am.Chem. Soc. 2005, 127, 3439.

(11) Zimmerman, J. L.; Bradley, R. K.; Huffman, C. B.; Hauge, R. H.;Margrave, J. L. Chem. Mater. 2000, 12, 1361.

(12) Star, A.; Joshi, V.; Skarupo, S.; Thomas, D.; Gabriel, J. C. P. J.Phys. Chem. B 2006, 110, 21014.

(13) Kong, J.; Dai, H. J. J. Phys. Chem. B 2001, 105, 2890.(14) Na, P. S.; Kim, H. J.; So, H. M.; Kong, K. J.; Chang, H. J.; Ryu,

B. H.; Choi, Y. M.; Lee, J. O.; Kim, B. K.; Kim, J. J.; Kim, J. H. Appl.Phys. Lett. 2005, 87, 093101.

(15) Monthioux, M.; Smith, B. W.; Burteaux, B.; Claye, A.; Fischer,J. E.; Luzzi, D. E. Carbon 2001, 39, 1251.

(16) Zhou, W.; Ooi, Y. H.; Russo, R.; Papanek, P.; Luzzi, D. E.; Fischer,J. E.; Bronikowski, M. J.; Willis, P. A.; Smalley, R. E. Chem. Phys. Lett.2001, 350, 6.

(17) Montoro, L. A.; Luengo, C. A.; Rosolen, J. M.; Cazzanelli, E.;Mariotto, G. Diamond Relat. Mater. 2003, 12, 846.

(18) Duesberg, G. S.; Muster, J.; Byrne, H. J.; Roth, S.; Burghard, M.Appl. Phys. A: Mater. Sci. Process. 1999, 69, 269.

(19) Wu, Z. C.; Chen, Z. H.; Du, X.; Logan, J. M.; Sippel, J.; Nikolou,M.; Kamaras, K.; Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.; Rinzler,A. G. Science 2004, 305, 1273.

(20) Zhang, Q.; Vichchulada, P.; lipscomb, L. D.; Lay, M. D. J. Mater.Sci. 2008, 44, 1206.

(21) Vichchulada, P.; Shim, J.; Lay Marcus, D. J. Phys. Chem. C 2008,112, 19186.

(22) Yu, A. P.; Bekyarova, E.; Itkis, M. E.; Fakhrutdinov, D.; Webster,R.; Haddon, R. C. J. Am. Chem. Soc. 2006, 128, 9902.

(23) Jia, H. B.; Lian, Y. F.; Ishitsuka, M. O.; Nakahodo, T.; Maeda, Y.;Tsuchiya, T.; Wakahara, T.; Akasaka, T. Sci. Technol. AdV. Mater. 2005,6, 571.

(24) O’Connell, M. J.; Bachilo, S. M.; Huffman, C. B.; Moore, V. C.;Strano, M. S.; Haroz, E. H.; Rialon, K. L.; Boul, P. J.; Noon, W. H.; Kittrell,C.; Ma, J. P.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E. Science 2002,297, 593.

(25) Vichchulada, P.; Shim, J.; Lay, M. D. J. Phys. Chem. C 2008, 112,19186.

(26) Vichchulada, P.; Zhang, Q.; Lay, M. D. Analyst 2007, 132, 719.(27) Pelletier, M. J., Ed.; Analytical Applications of Raman Spectroscopy;

Blackwell Science: Oxford, 1999.(28) Dresselhaus, M. S.; Dresselhaus, G.; Jorio, A.; Souza, A. G.; Saito,

R. Carbon 2002, 40, 2043.(29) Moon, J. M.; An, K. H.; Lee, Y. H.; Park, Y. S.; Bae, D. J.; Park,

G. S. J. Phys. Chem. B 2001, 105, 5677.(30) Furtado, C. A.; Kim, U. J.; Gutierrez, H. R.; Pan, L.; Dickey, E. C.;

Eklund, P. C. J. Am. Chem. Soc. 2004, 126, 6095.(31) Ma, J.; Wang, J. N. Chem. Mater. 2008, 20, 2895.(32) Kim, U. J.; Furtado, C. A.; Liu, X. M.; Chen, G. G.; Eklund, P. C.

J. Am. Chem. Soc. 2005, 127, 15437.(33) Itkis, M. E.; Perea, D. E.; Niyogi, S.; Rickard, S. M.; Hamon, M. A.;

Zhao, B.; Haddon, R. C. Nano Lett. 2003, 3, 309.(34) Bekyarova, E.; Itkis, M. E.; Cabrera, N.; Zhao, B.; Yu, A. P.; Gao,

J. B.; Haddon, R. C. J. Am. Chem. Soc. 2005, 127, 5990.

JP9086738

Figure 5. Plot of absorbance at 750 nm vs number of centrifugationcycles for an AP-grade SWNT soot suspension with a startingconcentration of 1 mg/mL of SWNT soot. The absorbances decreasefrom 2.576 to 0.175 as the number of centrifugation cycles increases.

Nondestructive Purification Method for SWNT Suspensions J. Phys. Chem. C, Vol. 114, No. 1, 2010 657