Characterization of Carbon Nanotubes by Raman spectroscopy

S. Costa1, E. Borowiak-Palen*1, M. Kruszyska1, A. Bachmatiuk1, R J. Kaleczuk1

1) Centre of Knowledge Based Nanomaterials and Technologies, Institute of Chemical and Environment Engineering, Szczecin University of Technology, Poland

*Corresponding author: email: eborowiak@ps.pl

Abstract
In current work we present the application of Raman spectroscopy to analyze carbon nanotubes. We show that having a cocktail of different carbon nanotubes samples one can easily distinguish in a quick measurement the presence of singlewalled, doublewalled and multiwalled carbon nanotubes (SWCNT, DWCNT, MWCNT, respectively). The so-called G-line is a characteristic feature of the graphitic layers and corresponds to the tangential vibration of the carbon atoms. The second characteristic mode is a typical sign for defective graphitic structures (D-line). The comparison of the ratios of these two peaks intensities gives a measure of the quality of the bulk samples. In addition, there is a third mode, named the radial breathing mode (RBM) which is very sensitive to the diameter of SWCNT and DWCNT. Additional option is the application of Raman microscopy for mapping analysis and depth profiling to view the intensity change in different direction of the sample.

Keywords Raman mapping ; diameter of carbon nanotubes .

Introduction
Carbon nanotubes (CNT) since their discovery become a important scientific field for the extensive research due to their interesting physics properties and technological applications. CNT have proven to be an unique system for the study of Raman spectra in one-dimensional (1D) systems [1, 2], and at the same time Raman spectroscopy has provided an exceedingly powerful tool useful in the study of the vibrational properties and electronic structures of CNT [3, 4], particularly for characterization of CNTs with respect to their diameters, and quality of the samples properties [5, 6]. By Raman spectroscopy different carbon materials can be analysed including singlewall, doublewall and multiwall carbon nanotubes, but unfortunately quantitative determination of each type are not possible at present state of the art [7].

Features of Raman spectroscopy SWCNT &DWCNT The Raman spectra presents different features being all sensitive to (n,m), such as the radial breathing mode (RBM), where all the carbon atoms are moving in-phase in the radial direction, the G-band where neighbouring atoms are moving in opposite directions along the surface of the tube as in 2D graphite, the

dispersive disorder induced D-band and its second-order related harmonic G-band. From these four features the RBM is the one who shows to be more sensitive to the nanotube diameter (dt) [8], according to the expression

RBM =

A +B dt

(1)

where A and B are constants and vary between individual tubes and bundle tubes [1, 9]. Some authors consider only the constant A in the diameter determination [6, 10, 11]. A doublewall carbon nanotube (DWCNT) can be considered to be a kind of MWCNT for which the interlayer interaction is generally considered to be turbostratic between the inner and outer nanotubes. For armchairarmchair DWCNT, some commensurate structure can be expected and the splitting of the Gband which is observed in 3D graphite could be seen. Another novel direction for future exploration is the small RBM linewidths (down to 0.25 cm1) occurring for the inner wall tube within an isolated DWCNT [1].

By measurement of RBM for many laser energies, the diameter distribution of the nanotubes in a particular SWCNT bundle can be found [11]. The G-band is thus an intrinsic feature of carbon nanotubes that is closely related to vibrations in all sp2 carbon materials. The most important aspect of the G-band is the characteristic Raman lineshape which differs in accordance with whether the nanotube is semiconducting or metallic, allowing to readily distinguishing between both types. This band shows two different components, the lower frequency component associated with vibrations along the circumferential direction, G-), and the higher frequency component, (G+), attributed to vibrations along G the direction of the nanotube axis. Previous studies show that the first component is dependent of diameter of nanotube while the second, the higher frequency component, doesnt present this dependence, which differs according to metallic or semiconducting nanotubes [10]. The D-band and G-band features are both observed in the Raman spectra of semiconducting and metallic SWCNT at the single nanotube level. The D-band in graphite involves scattering from a defect which breaks the basic symmetry of the graphene sheet, it is observed in sp2 carbons containing porous, impurities or other symmetry-breaking defects. On the other hand, the second-order G-band doesnt require an elastic defect-related scattering process, and is observable for defect-free sp2 carbons. These bands show a dependence of chirality and diameter of nanotubes [1] and energy laser [12]. The analysis of all resonance Raman effects has been greatly aided by the introduction of Kataura plot, of the interband transitions, Eii, as a function of dt for all values of (n,m) showing that each pair of indices in a nanotube has a unique set of Eii transition energies, which is physically due to the trigonal warping effect of the constant energy contours for a graphite sheet [11].

Multiwalled carbon nanotubes Multi-walled carbon nanotubes (MWCNT) are made of concentric graphene sheets rolled in a cylindrical form with diameters of tens of nanometers [13] and because of this large diameter of the outer tubes for typical MWCNT and because they contain an ensemble of carbon nanotubes with diameters ranging from small to very large, most of the characteristic differences that distinguish the Raman spectra in SWCNT

from the spectra for graphite are not so evident in MWCNT. For example, the RBM Raman feature associated with a small diameter inner tube (less than 2 nm) can sometimes be observed when a good resonance condition is established, but this is not the usual result, since the RBM signal from large diameter tubes is usually too weak to be observable and the ensemble average of inner tube diameter broadens the signal [1]. Whereas the G+ G splitting is large for small diameter SWCNT tubes, the corresponding splitting of the G band in MWCNT is both small in intensity and smeared out due to the effect of the diameter distribution within the individual MWCNT, and because of the variation between different tubes in an ensemble of MWCNT in typical experimental samples. Therefore the G-band feature predominantly exhibits a weakly asymmetric characteristic lineshape, with a peak appearing close to the graphite frequency. The most effective radiation influence is expected in the case of the electron irradiation of MWCNT. This is explained by the possibility not only that radiation defects occur during irradiation, which will contribute to the degradation of nanotubes, but also that seams appear due to the broken connections as vacancies are generated [14]. In the present study we present a simple way of evaluation of different physical properties of SWCNT from Raman spectroscopy analysis.

Experiment
The synthesis of SWCNT was described elsewhere [15] and the Raman analysis were carried out using a micro-Raman Renishaw spectrometer equipped with a CCD detector. Two laser lines were used to excite the sample, 514 nm (green laser) and 785 nm (red laser). Depth profile and mapping of a select area was also obtained with Raman spectrometer. All the measurements were performed at room

temperature. Results and Discussion

Comparison between green and red lasers The response for excitation with different lasers revealed very rich and strongly differing spectra for each sample [16-18]. Fig.1a presents two examples, one for excitation with the green laser (514nm) and other for excitation with the red laser (785nm). The resonance scattering caused by different lasers give us information about different excitations levels as show in Kataura plot (Fig. 2). In the same figure it is marked the energy separation of the lasers used in our measurements. In order to gain more information about the type of nanotubes excited by them it is necessary to calculate the diameter distribution, using data from red laser, since RBM with green laser is not clear. Also the intensity of G band increase when we use the 785nm laser.

785 nm 514 nm

*20
0,9 1,0 1,1 1,2 1,3 1,4

150

300

1500
-1

2000

2500

Diameter distribuition (nm)

Raman shift [cm ]

Fig. 1- (a) Different Raman spectra of SWCNT obtained from red laser (785 nm) and green laser (514 nm), (b) and its diameter distribution.

semiconducting tubes metallic tubes ___ green laser ___ red laser

Fig. 2- Kataura plot showing the different resonance excitations as function of different energy lasers and diameters. Determination of diameter distribution In the Fig.1a we present a Raman spectra where can be seen the four features of a SWCNT sample. An RBM band is visible in the red laser spectra. For typical SWCNT bundles in the diameter range dt = 1.5 0.2nm, A=234 cm1 nm and B=10 cm1 has been found [1], where B is an upshift in RBM due to tubetube interactions. The calculated diameter distribution was obtained using expression (1) and is presented on Table 1 and Fig 1b. The range of diameters vary between 0,8 nm and 1,5 nm. According to Kataura plot with the red laser we only obtain signal from semiconducting tubes but the green laser show that we also have metallic tubes. These facts indicate how important is the analysis with different energy lasers, because each one can give different optical and physical properties. The RBM band also allows distinguishing between SWCNT, DWCNT and MWCNT. In the Fig.3a is shown a spectrum comparing

the different RBM bands in nanotubes. DWCNT spectrum presents a double RBM due to the diameters of the inner and outer tubes. According to the expression (1) the band of higher Raman shift corresponds to the inner diameter and the one with less Raman shift to the outer diameter. Table 1- Diameter distribution calculated from RBM feature, using expression (1).

Raman shift / cm-1

152,275 183,419 203,352 214,397 225,421 234,225 267,121

Diameter / nm
1,441997 1,209809 1,096779 1,042795 0,993964 0,958133 0,844396

D band - G band relation The ratio between the D band and G band is a good indicator of the quality on bulk samples. If these both bands have similar intensity this indicates a high quantity of structural defects. In the Fig.3b is presented this relation in different kind of nanotubes. MWCNT spectrum is the one who shows the lowest ration, consequently higher quantity of structural defects due to its multiple graphite layers. Both SWCNT and DWCNT show a higher difference between D and G band. The ratio between D and G band and the RBM and its relation with diameter distribution, are very important factors in the way that allows us to distinguish between the three variants of nanotubes with one single analysis, which is a probe of the high performance of Raman spectrometer.

SWCNT DWCNT MWCNT

Intensity [arb. u.]

Intensity [arb. u.]

1200

150

200

250

300
-1

350

1350

1500

1650
-1

Raman shift [cm ]

Raman shift [cm ]

Fig. 3- Different Raman spectra obtained from SWCNT, DWCNT and MWCNT (a) RBM and (b) D and G bands (laser excitation 785 nm)..

Mapping
The first electron transition energy between semiconducting SWNTs is usually too small to be observed with standard Raman spectroscopy setups. However it is possible to observe it in a convenient range in a multi dimensions map [1]. A map of each feature can be obtained with this technique. The mapping allows to observe the homogeneity of the sample by screening the bigger area as in case of the standard Raman microscope measurement [19]. In the Fig. 4 is presented a mapping image of the three most important features in Raman spectrum, and its distribution in the area of the analyzed sample. This distribution can be easily observed with the different colors of the graphs. The mapping spectra can be visualized in two or three dimensions with similar results such as the images presented in the Fig. 5.
Raman s hift / cm-1

5000

4000
20 m

20 m

3000 Counts

2000

1000

500

1000

1500 Raman shift / cm-1

2000

2500

3000

RBM
Raman shift / cm-1

Raman shift / cm-1

G band

D band

Fig. 4- Raman two dimensional map images of RBM, G band and D band, in the selected area show in the inset (laser excitation 785 nm).

Fig. 5- Raman mapping of RBM band. Comparison between a top view (upper panel) and a 3D view (bottom panel). The colors (see right side) indicates the intensity of Raman intensity (laser excitation 785 nm).

Depth profile With this technique it is also possible to analyse the Raman response as a function of the depth of the sample. The depth profile of G-band of SWCNT is show in the Fig. 6 and it allows choosing the area of the sample which would give the bigger intensity of Raman signal. In the present case one can conclude that the best signal is obtained at 1m depth in the sample. The depth profile can also be visualized in two (Fig. 6a) or three dimensions (Fig. 6b).

350 0 340 0 330 0 320 0 310 0 -4 -2 0 2 4

350 0 300 0 250 0 200 0 150 0 100 0 50 00 50 0 100 0

Depth / m

a
150 200 0Raman shift 0
1 / cm0

250

300 0

Fig. 6- Raman depth profile of G band in SWCNT (a) in two dimensions and (b) in three dimensions

Conclusions
Raman spectroscopy is a powerful technique with different options such as area mapping, depth profile for beyond the standard spectra analysis. The energy of excitation laser is also very important due to the different results obtained, as show in Kataura plot. One can also easily determine the presence of different sorts of carbon nanotubes in the analyzed sample including determination of the diameter in case of SWCNT and MWCNT. The G/D ratio is a direct measure of the quality of the sample. Mapping analysis is important to see the homogeneity of the sample and compare the different Raman features in the select area. The depth profile is useful beyond the standard spectra, to select how the Raman signal changes in respect of the depth of the sample.

References
[1] M.S. DRESSELHAUS, G. DRESSELHAUS, R. SAITO, A. JORIO, Physics Reports, 409 (2005), 47-99 [2] A. JORIO, M.A. PIMENTA, C. FANTINI, M. SOUZA, A.G. SOUZA FILHO, GE.G. SAMSONIDZE, G. DRESSELHAUS, M.S. DRESSELHAUS, R. SAITO, Carbon, 42 (2004), 10671069 [3] L. KAVAN, P. RAPTA, L. DUNSCH, M. J. BRONIKOWSKI, P. WILLIS, R. E. SMALLEY, J. Phys. Chem. B, 105 (2001), 10764-10771 [4] L. ALVAREZ, A. RIGHI, T. GUILLARD, S. ROLS, E. ANGLARET, D. LAPLAZE, JEAN-LOUIS SAUVAJOL, Chemical Physics Letters, 316 (2000), 186190

[5] T. YU. ASTAKHOVA, G. A. VINOGRADOV, M. MENON, Russian Chemical Bulletin, International Edition, 52 (4) (2003), 823-830 [6] M. S. DRESSELHAUS, G. DRESSELHAUS, A. JORIO, A. G. SOUZA FILHO, M. A. PIMENTA, D R. SAITO, Acc. Chem. Res., 35 (2002), 1070-1078 [7] W. QIAN, T. LIU, F. WEI, H. YUAN, Carbon, 41 (2003), 1851-1864 [8] N. TAKEDA, K. MURAKOSHI, Anal Bioanal Chem., 388 (2007), 103108 [9] S.K. DOORN, D.A. HELLER, P.W. BARONE, M.L. USREY, M.S. STRANO, Appl. Phys. A, 78 (2004), 11471155 [10] M.S. DRESSELHAUS, G. DRESSELHAUS, A. JORIO, A.G. SOUZA FILHO, R. SAITO, Carbon, 40 (2002), 20432061 [11] M.S. DRESSELHAUS, A. JORIO, A.G. SOUZA FILHO, G. DRESSELHAUS, R. SAITO, Physica B, 323 (2002), 1520 [12] H.B. ZHANG, G.D. LIN, Z.H. ZHOU, X. DONG, T. CHEN, Carbon, 40 (2002), 24292436 [13] E.F. ANTUNES, A.O. LOBO, E.J. CORAT, V.J. TRAVA-AIROLDI, A.A. MARTIN, C. VERISSIMO, Carbon, 44 (2006), 22022211 [14] U. RITTER, P. SCHARFF, C. SIEGMUND, O.P. DMYTRENKO, N.P. KULISH, YU.I. PRYLUTSKYY, N.M. BELYI, V.A. GUBANOV, L.I. KOMAROVA, S.V. LIZUNOVA, V.G. POROSHIN, V.V. SHLAPATSKAYA, H. BERNAS, Carbon, 44 (2006), 26942700 [15] D. GRIMM, A. GRUNEIS, C. KRAMBERGER, M. RUMMELI, T. GEMMING, B. BUCHNER, A. BARREIRO, H. KUZMANY, R. PFEIFFER, T. PICHLER, Chemical Physics Letters 428 (2006) 416420 [16] S.K. DOORN, M.J. OCONNELL, L. ZHENG, Y.T. ZHU, S. HUANG, J. LIU, Physical review letters, 16802 (2005), 1-4 [17] H. KUZMANY, B. BURGER, A. THESS, R.E. SMALLEY, Carbon, 36 (5-6) (1998), 709-712 [18] J. KURTI, H. KUZMANY, B. BURGER, M. HULMAN, J. WINTERB, G. KRESSE, Synthetic Metals, 103 (1999), 2508-2509 [19] H. KUZMANY, W. PLANK, M. HULMAN, C. KRAMBERGER, A. GRUNEIS, T. PICHLER, H. PETERLIK, H. KATAURA, Y. ACHIBA, Eur. Phys. J. B, 22 (2001), 307-320

List of figure captions

Fig. 1- (a) Different Raman spectra of SWCNT obtained from red laser (785 nm) and green laser (514 nm), (b) and its diameter distribution. Fig. 2- Kataura plot showing the different resonance excitations as function of different energy lasers and diameters. Fig. 3- Different Raman spectra obtained from SWCNT, DWCNT and MWCNT (a) RBM and (b) D and G bands. Fig. 4- Raman two dimensional map images of RBM, G band and D band, in the selected area show in the inset. Fig. 5- Raman mapping of RBM band. Comparison between a top view (upper panel) and a 3D view (bottom panel). The colors (see right side) indicates the intensity of Raman intensity. Fig. 6- Raman depth profile of G band in SWCNT (a) in two dimensions and (b) in three dimensions