Evaluation of the individualization state in single-walled carbon nanotube solutions using

absorption, Raman and photoluminescence spectroscopy

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2012 Meas. Sci. Technol. 23 125501

(http://iopscience.iop.org/0957-0233/23/12/125501)

Download details:

IP Address: 93.180.53.211

The article was downloaded on 17/09/2013 at 06:35

Please note that terms and conditions apply.

View the table of contents for this issue, or go to the journal homepage for more

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING MEASUREMENT SCIENCE AND TECHNOLOGY

Meas. Sci. Technol. 23 (2012) 125501 (9pp) doi:10.1088/0957-0233/23/12/125501

Evaluation of the individualization statein single-walled carbon nanotubesolutions using absorption, Raman andphotoluminescence spectroscopyWonsuk Jung1, Ju Yeon Woo2, Seung Ho Lee1, Duckjong Kim3,Soohyun Kim1 and Chang-Soo Han2

1 Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology,373-1 Guseong, Yuseong, Daejeon 305806, Korea2 School of Mechanical Engineering, Korea University Anam, Seongbuk, Seoul 136-701, Korea3 Department of Nano-Mechanics, Korea Institute of Machinery & Materials, 171 Jang, Yuseong,Daejeon 305343, Korea

E-mail: cshan@korea.ac.kr

Received 19 March 2012, in final form 5 September 2012Published 6 November 2012Online at stacks.iop.org/MST/23/125501

AbstractWe report a method for characterizing the individualization of single-walled carbon nanotubes(SWCNTs) in a solution based on absorption, Raman and photoluminescence spectroscopy. Inthe first van Hope transition interband (S11), we selected two distinct peaks and introducedkurtosis as a quantitative value for representing the sharpness and magnitude of the peaks.Three samples of individualized SWCNTs prepared using our condition and two publishedconditions were analyzed using kurtosis. In addition, we used the radial breathing mode peaksand photoluminescence map to estimate the individualization qualitatively. As the result, wefound that three methods to characterize the individualization state showed consistent resultswith each other and could be effectively utilized for making a well-individualized sample.

Keywords: single-walled carbon nanotube, individualization, fluorescence, Raman, absorptionspectroscopy

(Some figures may appear in colour only in the online journal)

1. Introduction

Since the discovery of carbon nanotubes in 1991 [1], theyhave been researched in many fields of science [2, 3] dueto their excellent optical [4, 5], electrical [6] and structuralproperties [7–9]. Single-walled carbon nanotubes (SWCNTs)are outstanding candidates for applications in electronicdevices, transparent conductive films [10] and sensors [10, 11].

These applications require well-dispersed SWCNTs insolution. Moreover, the individualization of SWCNTs isrequired for devices that rely on the properties of an individualSWCNT. For example, in order to make a field-effecttransistor with an individual SWCNT, we should individualizeSWCNTs in the bundled state. In addition, SWCNTs should be

individualized before separating them according to diameterand chirality [12–17].

There are several methods for individualizing SWCNTs insolution, including the use of dimethylformamide, conjugatedpolymers, DNA and various surfactants [12–17]. Of these,surfactants like sodium dodecyl sulfide (SDS) are widelyused for dispersion and individualization. During SWCNTdispersion with surfactants, sonication is necessary forseparating bundled SWCNTs in the solution. In addition,ultracentrifugation is used to remove any remaining bundledor aggregated SWCNTs and impurities such as carbon debrisand metal catalysts. Previous studies have mostly used opticalmethods to estimate the purity and dispersion of SWCNTs.Absorption spectroscopy in the visible to near-infrared range

0957-0233/12/125501+09$33.00 1 © 2012 IOP Publishing Ltd Printed in the UK & the USA

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

(a)

(b) (c)

Figure 1. A schematic illustration of the three conditions used for individualizing SWCNTs. (a) C1 is our proposed method, which used ashorter ultrasonication time and slower centrifugation rate than other methods. (b) C2 puts about 5400 J of energy into the solution duringsonication and involves centrifugation at 63 000 rpm for 4 h. (c) C3 is a quick process, including centrifugation for 14 min.

is the optical tool used most widely for analyzing SWCNTs[4, 5, 18, 19].

However, the estimation of the degree of individualizationfor multiple individualized samples has not been welldocumented. Atomic force microscopy can be used to takeimages of individualized SWCNTs on a substrate, but thisproduces artifacts because of problems during the sampletreatment. This approach is not appropriate for estimatinga large quantity of individualized SWCNTs dissolved inaqueous solution. Alternatively, the fluorescence peaks inabsorption spectra have been used to estimate the state ofthe individualization, because individualized semiconductingSWCNTs have distinct absorption peaks in the first van Hopetransition band (S11). However, if two samples have similarlyshaped fluorescence peaks, it is difficult to determine whichone is more individualized, because such comparisons are stillvery qualitative. Therefore, it is necessary to find a quantitativeestimation method for efficient individualization.

This paper proposes a quantitative approach based onabsorption spectroscopy for estimating the individualizationstate of SWCNTs. We show that the radial breathing mode(RBM) in 514, 633 and 785 nm Raman spectra, fluorescencepeaks in the 785 nm Raman spectrum and three-dimensional(3D) photoluminescence (PL) images can be used forestimating the individualization of SWCNTs with consistentresults. In addition, we show that our proposed condition forindividualization is more efficient than previous conditions[20, 21] in terms of the quality and yield of individualizedSWCNTs.

2. Experimental methods

2.1. Materials

In order to individualize SWCNTs, ultrasonication andcentrifugation are usually used [16]. Before individualization,

SWCNTs typically existed in bundled states due to van derWaals attraction to each other. Cathode-type surfactants, whichcause steric hindrance and static electrical repulsive forces, areneeded to isolate a SWCNT in an individualized state [22–24].Here, we prepared samples using both reported conditions andour new condition.

2.2. Three conditions for preparing SWCNT samples

We investigated the individualization process using threedifferent conditions to prepare samples (figure 1). Ourproposed method is condition 1 (C1). This uses pure watercontaining 1 wt% aqueous SDS, e.g., 50 mL of deionized(DI) water containing 0.5 g of SDS. First, SDS was mixedin DI water by ultrasonication at 30 W for 10 min. Then,HiPco SWCNTs (15 mg) were added to the SDS solution,which was sonicated further. Bundled SWCNTs were removedby ultracentrifugation for 3 h in an Optima MAX (BeckmanCoulter), as shown in figure 1. Finally, 40% of the supernatantwas decanted. Condition 2 (C2) is a widely used conditionfor individualizing SWCNTs in 1 wt% aqueous SDS/H2Osolution, e.g., 100 mL of DI water and 1 g of SDS. This solutionwas sonicated for 72 min at 75 W to apply 5400 J. Then, thesolution was centrifuged at 63 000 rpm for 4 h [20], and 40% ofthe supernatant of the solution was decanted. Condition 3 (C3)uses 2 wt% aqueous SDS/H2O solution and ultrasonicationfor 60 min at 36 W. The solution was centrifuged for a shortertime, e.g., 14 min at 54 000 rpm [21]. As with C1 and C2, 40%of the supernatant was decanted. Comparing the conditionsused for the three samples, C1 involves less time and energythan C2 and C3, i.e. 40 min and 3000 J. In addition, C1 usedcentrifugation at the lowest speed, 54 000 rpm.

2.3. Optical measurements

To characterize the individualization state of SWCNTs,we used three optical methods: absorbance spectroscopy,

2

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

Figure 2. The absorbance spectra show differences in the peaksdepending on the three conditions used. Sharp peaks appear in thefirst van Hove transition region 800–1400 nm, especially forwavelengths >1000 nm.

Raman spectroscopy and PL maps. We investigated theabsorbance spectra from 400 to 1400 nm using a UV-VIS-NIR Spectrophotometer (UV-3600, Shimadzu). The Ramanspectra were obtained from the RBM to the fluorescence peaksusing Raman spectroscopy (RE02 InVia Raman Microscope,Renishaw) at laser wavelengths of 514, 633 and 785 nm. Inaddition, a PL apparatus (Fluorolog PL3D, Jobin Yvon Horiba)using emission wavelengths from 550 to 750 nm and excitationfrom 900 to 1250 nm was used to analyze the individualizedsemiconducting SWCNTs.

3. Results and discussion

3.1. Analysis of absorption spectrum

The dispersion state of SWCNTs individualized in aqueoussolution can be analyzed using absorption spectra. In addition,we can estimate the relative portion of semiconducting andmetallic nanotubes [5, 22, 25]. As SWCNT individualizationincreases, the absorption peaks typically become sharper andmore distinct. When the SWCNTs are in bundled and ropedstates, the peaks in the spectra are small and broad [18, 26–28].In particular, peak broadening and red shifts in the S11 band(800–1400 nm) are affected by the degree of individualization[5, 18].

The absorption spectra of samples prepared using C1, C2and C3 had distinct peaks in the S11 range as shown in figure 2.Therefore, we considered these samples to be individualized.However, compared with C3, the spectra of samples preparedusing C1 and C2 had sharper peaks over the entire wavelengthrange. Two large, sharp peaks (at about 1150 and 1260 nm) inthe S11 range seemed to be representative; therefore, we usedthese peaks to estimate the individualization state. Comparingthe three spectra qualitatively, the solutions prepared using C1and C2 had sharper, larger peaks at 1150 and 1260 nm than thesolution prepared using C3. This indicates greater dispersionof isolated SWCNTs with specific chirality, which can beused to estimate SWCNT individualization. Moreover, sample

C3 contained some noise at 400–500 and 800–900 nm. Thisphenomenon results from SWCNTs aggregated in bundles orroped states via van der Waals forces [22, 29]. The aggregationstates arise from the transfer of energy to adjacent SWCNTsthat have smaller bandgaps. By contrast, the spectra of C1 andC2 were similar in terms of peak sharpness and magnitude.Consequently, we need to quantify the breadth and magnitudeof the peaks in the spectra, because these parameters areaffected by the degree of individualization.

Hence, we developed a quantitative formula based ona statistical concept, kurtosis, for simultaneously estimatingthe breadth and peak-to-valley ratio of the distinct peaksin the S11 region. This concept of kurtosis can distinguishthe performances of C1 and C2 by estimating the degreeof individualization. Kurtosis is a mathematical method foranalyzing the sharpness and magnitude of data [30, 31]. Thismethod considers the absolute values of the height and breadthof a peak. Higher values of kurtosis indicate a higher andsharper peak, while lower values indicate a lower and lessdistinct peak. These properties can be applied to evaluate theindividualization data of SWCNTs.

Kurtosis =[

N(N + 1)

(N − 1)(N − 2)(N − 3)×

∑ (X − X̄

s

)4]

(1)

Normalized kurtosis = Kurtosis of sample

Absorbance at 633 nm(2)

N: number of dataX : value of each datumX̄ : mean value of datas: standard deviation of data N.

Kurtosis reflects the magnitude, mean values, standarddeviation and number of samples, as shown in equation (1).The data measured are the absorbance obtained at eachwavelength in the spectrum. We calculated the kurtosis ofthe two large peaks at 1150 and 1260 nm for C1 and C2. Therange of the samples is from left to right for the valleys in thelarge peaks. Finally, we normalized the kurtosis value usingthe absorbance at 633 nm to remove the effect of the solutionconcentration.

Considering the quantitative result using kurtosis, C1appeared to be the best condition for individualizing SWCNTsamong three samples. The normalized kurtosis of C1 was2.44 (peak 1) and 1.93 (peak 2). These values are largerthan the values for C2 (peak 1, 2.14; peak 2, 1.71) andC3 (peak 1, 1.59; peak 2, 1.62), as shown in table 1. Ingeneral, each sample solution contained different quantitiesof nanotubes with specific chiralities. This likely resulted inthe difference in the magnitudes of the two distinct peaks.Each sample of SWCNT solution can have different dominantwavelengths at these two peaks. Therefore, to improve thereliability of the estimation, we added two kurtosis values. C1had larger kurtosis values than C2 and C3, indicating that thebest individualization performance was achieved using C1,as shown in table 1. This new mathematical approach is aquantitative method for comparing the individualization ofSWCNTs. In addition, this analysis method gave consistentestimates, as do qualitative methods like Raman spectroscopyand PL.

3

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

Table 1. Quantitative measurement using the kurtosis of the twohighest UV absorbance peaks in the first van Hove transition regionand the sums of the kurtosis of each peak.

Criterion of C1 C2 C3

Peak 1 Peak value (au) 0.4467 0.6811 0.7429Peak position (nm) 1130 1126 1110Normalized kurtosis 2.4425 2.1438 1.5943

Peak 2 Peak value (au) 0.5267 0.7452 0.6246Peak position (nm) 1268 1262 1252Normalized kurtosis 1.9265 1.7068 1.6178Summation of 4.3691 3.8505 3.2121normalized kurtosis

3.2. RBM analysis of Raman spectrum

Raman spectra have been used to analyze the diameter,chirality, purity, structure and state of individualization ofSWCNTs [22, 32, 33]. The Raman spectral peaks, includingthe D-peak, G-peak, RBM and fluorescence peaks, containmeaningful data on SWCNTs [2, 5]. In this study, weinvestigated the individualized state of SWCNTs based onpeaks in the RBM and fluorescence range. The peaks in theRBM range produced by excitation at 514, 633 and 785 nm haddifferent shapes according to the individualization state [34].The RBM peaks of the Raman spectra at each wavelengthreflect differences in the chirality of the SWCNTs.

First, excitation at 785 nm is typically used to investigatethe state of aggregation by observing the resonance modes ofsemiconducting SWCNTs [11, 33]. As shown in figure 3(a),the peak intensity at 240 cm−1 decreased slightly from C1 toC3, while the peak intensity at 280 cm−1 increased, indicatingaggregated SWCNTs [33]. The peak at 280 cm−1 was redshifted with respect to the peak at 267 cm−1. This redshifting of the electronic transitions is on the basis of bundledand roped SWCNTs. Several theoretical studies predicted abundling-induced red shift in transition energies [29, 45–47]and suggested that bundling can induce as much as an 8–10%upshift in RBM frequency [48–50].

As shown in figure 3(a), for C1, the peak intensity at280 cm−1 was extremely small, about 5% of the peak for C3,while the ratio for C2/C3 is 64% at the roping peaks. Thepeak at 280 cm−1 shifted from 267 cm−1 is called ‘ropingpeak’ since O’Connell et al [35] studied this peak that wasobserved when SWCNTs were bundled. Of the three peaks,C1 has the smallest peak, meaning more individualization. Butthe difference between C2 and C3 peaks is not comparableto the results of the absorption spectrum. In conclusion, theroping peak at 280 cm−1 at 785 nm Raman data indicates avarying degree of individualization of SWCNTs, in the orderC1 > C2 > C3.

With 514 nm excitation, we can observe metallic HiPcoSWCNTs [32, 33]. It has been reported that although thesemiconducting region is off resonance in the individualizedstate, aggregated SWCNTs may produce the resonance peak[33]. In our experiments, we did not observe specific changesaccording to the method used, as shown in figure 3(b). Thismeans that the RBM peaks produced by the 514 nm lasercannot be used to estimate the individualization state.

(a)

(b)

(c)

Figure 3. The RBM of Raman spectra of SWCNTs using excitationat (a) 785, (b) 633 and (c) 514 nm.

For excitation at 633 nm, the peak intensity at190 cm−1 attributed to metallic nanotubes differed markedlyfor the different samples, while the peaks in thesemiconducting range were similar (figure 3(c)). Althoughthe intensity of the metallic peaks at 190 cm−1 is markedlygreater for samples with greater individualization (C1 > C2> C3), we do not know the exact relationship between thedegree of individualization and the peak intensity. We can sayonly that the portion of metallic nanotubes indicated by the190 cm−1 Raman shift increased markedly. Further research isneeded to clarify the reason for this. However, the RBM peakusing 633 nm excitation is not sufficient to characterize theindividualization of SWCNTs.

4

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

(a)

(b)

Figure 4. Picture of the G-peak of Raman spectra. (a) 633 nmRaman spectra. (b) 514 nm Raman spectra.

Table 2. FWHM and ratio of FWHM at the G-peak excited by the633 nm laser.

633 nm sample C1 C2 C3

G-band (cm−1) 1593.47 1589.93 1591.70FWHM (cm−1) 18.4827 18.8389 19.7535FWHM ratio (%) 93.56 95.37 100

3.3. Analysis of G-peak in 633 and 514 nm Raman spectra

Although the RBM analysis of Raman spectra of 514 and633 nm is not sufficient to evaluate the individualization ofSWCNTs, the sharpness of G-band can help us to estimatethe state of individualization of SWCNTs. In 633 nm Ramanspectra, though the whole shape of the G-band is similar(figure 4(a)), the sharpness of the G-peak is slightly different(table 2). The sharpness can be analyzed by the full-width athalf-maximum (FWHM) in which the G-peak is normalizedas an arbitrary unit. There is some tendency of the FWHM toincrease from C1 to C3. The FWHM of C1 is 18.4827 cm−1,which is smaller than the values for C2 (18.8389 cm−1) andC3 (19.7535 cm−1). Also, the ratios of the FWHM, of C1 toC3 and C2 to C3, are 93.56% and 95.37%.

This phenomenon arises at a G-peak of 514 nm Ramanspectra (figure 4(b)). The FWHM also increases from C1 toC3. The difference of the FWHM between C1 and C2 is lessthan the gap between C3 and them. The ratios of C1 and

(a)

(b)

Figure 5. Picture of the 785 nm Raman spectra. (a) Thefluorescence peak in 785 nm Raman spectra. (b) Roping andfluorescence peaks normalized using the G-peak.

Table 3. FWHM and ratio of FWHM at the G-peak excited by the514 nm laser.

514 nm sample C1 C2 C3

G-band (cm−1) 1592.17 1590.67 1592.17FWHM (cm−1) 24.2156 24.9895 76.9107FWHM ratio (%) 31.48 32.49 100

C2 to C3 are 31.48% and 32.49% as shown in table 3. Theseresults indicate that the best individualization performance wasachieved at C1.

3.4. Analysis of fluorescence peak in 785 nm Raman spectra

The broad fluorescence band of Raman spectra over2000 cm−1 is caused by isolated semiconducting SWCNTswith a bandgap. This fluorescence peak indicates the state ofindividualization and the presence of bundles of SWCNTs [5].Here, we measured the fluorescence peaks from Raman spectrausing 785 nm laser excitation. The peaks were normalizedusing the G-peak. Distinct fluorescence peaks are observedwhen nanotubes are well dispersed [4, 5, 36]. As shown infigure 5(a), C1 gave sharp, high peaks at 2587 and 2357 cm−1,while the peaks for C2 and C3 were broader and of lowerintensity. For comparison, we used the Raman data for the

5

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

(a) (b)

(c)

Figure 6. Pictures of the PL intensity of solutions prepared by each condition using emission wavelengths from 900 to 1250 nm andexcitation wavelengths from 500 to 750 nm. (a) Sample C1 has sharp, high peaks. (b) C2 has low peaks, but the peaks are clearlydistinguished. (c) C3 shows a noisy signal, and the peaks are not clearly distinguished. Moreover, the intensity of the peaks is low.

roping and fluorescence peaks, as shown in figure 5(b). All datawere normalized using the G-peak. The magnitudes of the twodistinct fluorescence peaks over 2000 cm−1 were summed. Thevalues for C1 and C2 were about 140% and 110% higher thanfor C3, respectively. In conclusion, C1 is the best method forindividualizing SWCNTs. As with the roping peak in 785 nmRaman spectra, the fluorescence peaks can be used to estimatethe degree of individualization qualitatively.

3.5. Analysis of the PL map

The PL map is used to characterize SWCNTs to determinethe distribution of the chiral indexes (n, m) of semiconductingSWCNTs [37–39]. Therefore, we postulated that the PL mapcould be used to estimate the degree of individualization,because we cannot obtain clear contours in the PL map unlessthe sample is well individualized [27, 35, 40]. The emissionintensity of the PL of an SWCNT sample is mostly affected bythe individualization state of the carbon nanotubes, while thePL intensity is not influenced by the specific wavelength oroptical properties of the solution [4]. In addition, the peakintensity of PL is dependent on the size of the SWCNTbundles and gives information on the aggregation dynamicsof SWCNTs [27, 41, 42].

Therefore, we measured the PL by scanning at excitationwavelengths from 500 to 750 nm and emission wavelengths

from 900 to 1250 nm. As shown in figure 6, C1 gave verysharp, clear peaks. These results were very similar to previousPL results for SWCNTs in 1% SDS [43]. By contrast, C2and C3 gave very small peaks and noise signals. Figure 7shows the 3D PL maps of the intensity and contour of theemissions, which imply the chirality of the SWCNTs. C1 hada maximum intensity of 8600 at chirality (8, 4) and had verybright, distinct contours at several chiralities, including (8, 4),(7, 5), (6, 5), (9, 4), (11, 0), (8, 3) and (8, 6), as shown infigure 7(a). The intensity of every peak exceeded 6000. Theabsorption spectrum of C1 showed many sharp peaks andspecific chiralities over all wavelengths from 400 to 1300 nm,indicating that the SWCNTs were well individualized. Bycontrast, the PL maps of C2 and C3 had broadened, lessbright contours, with reduced PL intensities. For C2, we candiscriminate the chirality at (9, 4), (11, 0), (7, 5), (7, 6), (8, 4)and (8, 3), although the intensity and contour are less clear thanfor C1. The highest intensity of C2 was 650 at (9, 4), whileall other chiralities for C2 had very weak intensities less than300. Finally, C3 gave a maximum intensity of 40 at (7, 5). Wecould not discriminate specific chiralities in C3. This resultmeans that C3 is insufficient for effective individualizationbecause the absorption spectrum of C3 had small, blunt peaksover the entire range and some noise at specific wavelengths.If SWCNTs are in the bundled or roped state, the emission

6

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

(a) (b)

(c )

Figure 7. PL maps describing the intensity and chirality of SWCNTs. (a) C1 has high intensity and SWCNTs of various chiralities. Thehighest intensity region is composed of (8, 4) and (9, 4). (b) C2 shows one-tenth the intensity of C1. Nevertheless, each peak can bedistinguished, and the highest region is (9, 4). (c) C3 has very weak intensity and shows only one kind of chirality (7, 5) because ofaggregation of the SWCNTs.

has a low intensity or noise signal that results from quenchingor shifting of energy due to an undesired energy transfer inPL. Moreover, effects from breadth of excitation wavelengthand not only emission wavelength should be studied in thefuture for an exact analysis of individualization of SWCNTs.Although PL maps especially show higher contrast of the peakintensities of C1, C2 and C3 than other optical methods, thesePL maps have the weakness that this method can be usedto evaluate only semiconducting and not metallic SWCNTs[44].

4. Conclusion

Using absorption spectroscopy, the RBM and fluorescencepeaks in Raman spectra and PL, we estimated theindividualization of dispersed SWCNT samples preparedusing three different conditions. Based on representativeabsorption peaks at the first van Hope transition interband(S11), we devised a quantitative method using the kurtosisvalue to estimate the individualization of the samples. Wedemonstrated that these results qualitatively matched theresults of the RBM and fluorescence peaks in 785 nm Ramanspectra, as well as the 3D PL map. Our new condition forindividualizing SWCNTs performed better than publishedconditions in terms of the quantitative and qualitative estimatesbased on these optical methods.

Acknowledgments

This research was supported by the Global Frontier ResearchCenter of Advanced Soft Electronics (MEST) and KoreaUniversity Grant.

References

[1] Iijima S 1991 Helical microtubules of graphitic carbon Nature354 56–8

[2] Dresselhaus M S, Dresselhaus G, Saito R and Jorio A 2005Raman spectroscopy of carbon nanotubes Phys. Rep.409 47–99

[3] Jorio A, Pimenta M A, Fantini C, Souza M, Souza Filho A G,Samsonidze G G, Dresselhaus G, Dresselhaus M Sand Saito R 2004 Advances in single nanotubespectroscopy: Raman spectra from cross-polarized light andchirality dependence of Raman frequencies Carbon42 1067–9

[4] Bachilo S M, Strano M S, Kittrell C, Hauge R H, Smalley R Eand Weisman R B 2002 Structure-assigned optical spectraof single-walled carbon nanotubes Science 298 2361–6

[5] O’Connell M J et al 2002 Band gap fluorescence fromindividual single-walled carbon nanotubes Science297 593–6

[6] Avouris P 2002 Molecular electronics with carbon nanotubesAcc. Chem. Res. 35 1026–34

[7] Roche S 2000 Mechanical and electronic properties of carbonnanotubes Ann. Chim. Sci. 25 529–32

[8] Schmldt G and Malwltz M M 2003 Properties ofpolymer–nanoparticle composites Curr. Opin. ColloidInterface Sci. 8 103–8

7

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

[9] Thostenson E T, Ren Z and Chou T W 2001 Advances in thescience and technology of carbon nanotubes and theircomposites: a review Compos. Sci. Technol.61 1899–912

[10] Shin D H, Song J W and Han C S 2009 Conductivity of filmsmade from single-walled carbon nanotubes in terms ofbundle diameters Scr. Mater. 60 607–10

[11] Yoon D K, Choi J B, Han C S, Kim Y J and Baik S H 2008The quantitative characterization of the dispersion state ofsingle-walled carbon nanotubes using Raman spectroscopyand atomic force microscopy Carbon 46 1530–4

[12] Zheng M, Jagota A, Semke E D, Diner B A, Mclean R S,Lustig S R, Richardson R E and Tassi N G 2003DNA-assisted dispersion and separation of carbonnanotubes Nature Mater. 2 338–42

[13] Chattopadhyay D, Galeska I and Papadimitrakopoulos F J2003 A route for bulk separation of semiconducting frommetallic single wall carbon nanotubes Am. Chem. Soc.125 3370–5

[14] Chen Z, Du X, Du M-H, Rancken C D, Cheng H-Pand Rinzler A G 2003 Bulk separative enrichment inmetallic or semiconducting single-walled carbon nanotubesNano Lett. 3 1245–9

[15] Kim D S, Nepal D and Geckeler K E 2005 Individualization ofsingle-wall carbon nanotubes: is the solvent important?Small 1 1117–24

[16] Moore V C, Strano M S, Haroz E H, Hauge R H, Smalley R E,Schmidt J and Talmon Y 2003 Individually suspendedsingle-walled carbon nanotubes in various surfactants NanoLett. 3 1379–82

[17] Dalton A B, Stephan C, Coleman J N, McCarthy B,Ajayan P M, Lefrant S, Bernier P, Blau W J and Byrne H J2000 Selective interaction of a semi-conjugated polymerwith single wall nanotubes J. Phys. Chem. B104 10012–6

[18] Naumov A V, Ghosh S, Tsyboulski D A, Bachilo S Mand Weisman R B 2011 Analyzing absorption backgroundsin single-walled carbon nanotube spectra ACS Nano5 1639–48

[19] Strano M S, Doorn S K, Haroz E, Kittrell C, Hauge R Hand Smalley R E 2003 Assignment of (n, m) Raman andoptical features of metallic single-walled carbon nanotubesNano Lett. 3 1091–6

[20] Krupke R, Hennrich F, von Lohneysen H and Kappes M M2003 Separation of metallic from semiconductingsingle-walled carbon nanotubes Science 301 344–7

[21] Arnold M S, Green A A, Hulvat J F, Stupp S Iand Hersam M C 2006 Sorting carbon nanotubes byelectronic structure using density differentiation NatureNanotechnol. 1 60–5

[22] Yoon D, Kang S-J, Choi J-B, Kim Y-J and Baik S 2007 Theevaluation of individual dispersion of single-walled carbonnanotubes using absorption and fluorescence spectroscopictechniques J. Nanosci. Nanotechnol. 7 3727–30

[23] Krstic V, Duesberg G S, Muster J, Burghard M and Roth S1998 Blodgett films of matrix-diluted single-walled carbonnanotubes Chem. Mater. 10 2338–40

[24] Duesberg G S, Muster J, Krstic V, Burghard M and Roth S1998 Chromatographic size separation of single-wallcarbon nanotubes Appl. Phys. A 67 117–9

[25] Zheng M et al 2003 Structure-based carbon nanotube sortingby sequence-dependent DNA assembly Science302 1545–8

[26] Itkis M E, Perea D E, Niyogi S, Rickard S M, Hamon M A,Hu H, Zhao B and Haddon R C 2003 Purity evaluation ofas-prepared single-walled carbon nanotube soot by use ofsolution-phase near-IR spectroscopy Nano Lett. 3 309–14

[27] Tan Y and Resasco D E 2005 Dispersion of single-walledcarbon nanotubes of narrow diameter distribution J. Phys.Chem. B 109 14454–60

[28] Makino K, Hirano A, Shiraki K, Maeda Y and Hase M 2005Purity assessment of SWNTs J. Phys. Chem. B109 9952–65

[29] Reich R and Thomsen C 2002 Electronic band structure ofisolated and bundled carbon nanotubes Phys. Rev. B65 155411

[30] Dodge Y 2003 The Oxford Dictionary of Statistical Terms(Oxford: Oxford University Press) ISBN 0-19-920613-9

[31] De Carlo L T 1997 On the meaning and use of kurtosisPsychol. Methods 2 292–307

[32] Kim W-J, Nair N, Lee C Y and Strano M S 2008 Covalentfunctionalization of single-walled carbon nanotubes alterstheir densities allowing electronic and other types ofseparation J. Phys. Chem. C 112 7326–31

[33] Heller D A, Barone P W, Swanson J P, Mayrhofer R Mand Strano M S 2004 Using Raman spectroscopy toelucidate the aggregation state of single-walled carbonnanotubes J. Phys. Chem. B 108 6905–9

[34] Thomsen C, Telg H, Maultzsch J and Reich S 2005 Chiralityassignments in carbon nanotubes based on resonant Ramanscattering Phys. Status Solidi b 242 1802–6

[35] O’Connell M J, Sivaram S and Doorn S K 2004 Near-infraredresonance Raman excitation profile studies of single-walledcarbon nanotube intertube interactions: a direct comparisonof bundled and individually dispersed HiPco nanotubesPhys. Rev. B 69 235415

[36] Shin D H, Kim J-E, Shim H C, Song J-W, Yoon J-H, Kim J,Jeong S, Kang J, Baik S and Han C-S 2008 Continuousextraction of highly pure metallic single-walled carbonnanotubes in a microfluidic channel Nano Lett.8 4380–5

[37] Weisman R B and Bachilo S M 2003 Dependence of opticaltransition energies on structure for single-walled carbonnanotubes in aqueous suspension: an empirical Kataura plotNano Lett. 3 1235–8

[38] Jones M, Engtrakul C, Metzger W K, Ellingson R J,Nozik A J, Heben M J and Rumbles G 2005 Analysis ofphotoluminescence from solubilized single-walled carbonnanotubes Phys. Rev. B 71 115426

[39] Miyauchi Y, Oba M and Maruyama S 2006 Cross-polarizedoptical absorption of single-walled nanotubes by polarizedphotoluminescence excitation spectroscopy Phys. Rev. B74 205440

[40] Hasan T, Scardaci V, Tan P H, Rozhin A G, Milne W Iand Ferrari A C 2008 Dispersibility and stabilityimprovement of unfunctionalized nanotubes in amidesolvents by polymer wrapping Physica E 40 2414–8

[41] Hasan T, Scardaci V, Tan P H, Rozhin A G, Milne W Iand Ferrari A C 2007 Stabilization and ‘debundling’ ofsingle-wall carbon nanotube dispersions inN-methyl-2-pyrrolidone (NMP) by polyvinylpyrrolidone(PVP) J. Phys. Chem. C 111 12594–602

[42] Tan P H, Rozhin A G, Hasan T, Hu P, Scardaci V, Milne W Iand Ferrari A C 2007 Photoluminescence spectroscopy ofcarbon nanotube bundles: evidence for exciton energytransfer Phys. Rev. Lett. 99 137402

[43] Wei L, Wang B, Goh T H, Li L-J, Yang Y, Chan-Park M Band Chen Y 2008 Selective enrichment of (6,5) and (8,3)single-walled carbon nanotubes via cosurfactant extractionfrom narrow (n,m) distribution samples J. Phys. Chem. B112 2771–4

[44] Lefebvre J, Maruyama S and Finnie P 2008Photoluminescence: science and applications Top. Appl.Phys. 111 287–319

8

Meas. Sci. Technol. 23 (2012) 125501 W Jung et al

[45] Kwon Y K, Saito S and Tomanek D 1998 Effect of intertubecoupling on the electronic structure of carbon nanotuberopes Phys. Rev. B 58 R13314–7

[46] Delaney P, Choi H J, Ihm J, Louie S G and Cohen M L 1999Broken symmetry and pseudogaps in ropes of carbonnanotubes Phys. Rev. B 60 7899–904

[47] Maarouf A A, Kane C L and Mele E J 2000 Electronicstructure of carbon nanotube ropes Phys. Rev. B 61 156–65

[48] Kuzmany H, Plank W, Hulman M, Kramberger C, Gruneis A,Pichler T, Peterlik H, Kataura H and Achiba Y 2001

Determination of SWCNT diameters from the Ramanresponse of the radial breathing mode Eur. Phys. J. B22 307–20

[49] Kukovecz A, Kramberger C, Georgakilas V,Prato M and Kuzmany H 2002 A detailed Ramanstudy on thin single-wall carbon nanotubes preparedby the HiPCO process Eur. Phys. J. B 28 223–30

[50] O’Connell M J 2006 Carbon Nanotubes Properties andApplications (London: Taylor and Francis) ISBN978-0-8493-2748-3

9