当前位置:首页 >> 理学 >>





Laboratoire de Physique Cristalline, Institut des Matériaux Jean Rouxel*, BP 32229, 44322 Nantes Cedex 03, France; b) National Institute of Materials Physics, Lab.160, Bucharest, PO Box MG-7, Romania


We present a survey of Raman and Surface Enhanced Raman Scattering (SERS) studies carried out on carbon nanotubes systems, including single walled nanotubes (SWNTs), multi-walled nanotubes (MWNTs) and nanotubes/polymers composites. We first recall the main features of Raman spectra taken at different excitation wavelengths and we particularly focus on the interactions that take place between tubes when they are in bundles. We predict an upshift of the radial breathing mode (RBM) and we demonstrate this effect experimentally. In the case of tubes deposited on a rough gold or silver layer, i.e. in SERS conditions, we show that interactions take place also with the metallic surface giving rise at some degradation of the tubes. Then, we present results on MWNTs and interpret the low frequency Raman modes, as originating from the RBM of isolated tubes whose interaction between the concentric shells lead again to an up-shift of the frequency of these modes. Finally, we have undertaken a careful study of composites prepared with nanotubes and either saturated or non-saturated polymers. For example, with Poly-(phenylene-vinylene)/nanotubes composites, it is shown that the luminescence properties of such compounds are strongly affected by the presence of the nanotubes. Raman Scattering, Surface Enhanced Raman Scattering, Carbon Nanotubes, Composites.

Key words:

E. C. Faulques et al. (eds.), Spectroscopy of Emerging Materials, 127--138. ? 2004 Kluwer Academic Publishers. Printed in the Netherlands.


S. Lefrant et al



Carbon nanotubes have been studied extensively since their first observation in 1991 by Ijima [1]. Several techniques of preparation have been developed such as laser ablation [2], arc electric procedure [3] or decomposition of hydrocarbides [4]. The tremendous interest of scientists towards this new class of carbon materials is due to several reasons. First, the electronic properties are of primary importance on a fundamental point of view, but more importantly, potential applications are proposed every day [5], far beyond the use of nanotubes as electron emitters in flat screen as reported some time ago. Let us recall that nanotubes can be considered as graphene sheets rolled up in different ways. If we consider the so-called chiral vectors c = na1 + na2 , in which a1 and a2 are the basis vectors of a 2D graphite lattice, depending on the value of the integers n and m, one can define three families of tubes : “armchair” tubes (n = m), “zig-zag” tubes (n or m = 0), and chiral tubes (n m 0). Band structure calculations have demonstrated that tubes are either metallic compounds, or zero-gap semiconductors, or semiconductors [6,7]. More commonly, they are divided into metallic tubes (when n-m is a multiple of 3) or semiconducting ones.



From the early days of investigation of carbon nanotubes, two spectroscopic techniques emerged as needed tools, namely high resolution transmission electron microscopy (HRTEM), and Raman scattering [3]. On one hand, HRTEM permits a direct observation of single-walled nanotubes (SWNTs) either isolated or in bundles providing then an estimation of the average diameter of the tubes. In the case of multiwalled nanotubes (MWNTs) HRTEM allows a determination of the size of the inner and outer diameters, as well as the averaged number of shells. As it will be seen later, these parameters turn out to be important for the simulation of Raman spectra for example. On the other hand, Raman scattering, via the vibrations, provides additional information, especially when performed at different wavelength excitations (Fig.1). Three domains of frequencies can be distinguished: i) the 1400-1700 cm-1 range where the so-called “G” modes around1600 cm-1 are seen, originating from a graphite vibration which is splitted as a consequence of the curvature of the tubes, ii) the frequency domain ranging from 1100 to 1400 cm-1 in which the “D” band is observed,

Raman and SERS Studies of Carbon Nanotubes


as a signature of defects, iii) the low frequency range in which appear the radial breathing modes of the tubes.

Figure 1. Raman spectra of single-walled carbon nanotubes recorded at room temperature for different excitation wavelengths: 457.9 nm; 514.5 nm; 676.4 nm and 1064 nm, as indicated on the curves.

The theoretical prediction of these vibrations was made by several authors (see for example refs. 8,9). Two main characteristics in the Raman spectra of SWNTs can be raised. First, the radial breathing mode, hereafter referred to RBM, is strongly dependant upon the diameter of the tubes.


S. Lefrant et al

Performing Raman scattering in high resolution in this domain reveals several components and several models have been built [10,11,12] to establish a direct relationship between the frequency of this mode and the tube diameter. In our group, we use the expression (cm-1) = 2238/d(?) as published by J.P. Buisson et al [13]. As a consequence, one can determine some kind of diameter distribution of tubes in a sample, if one keeps the laser excitation unchanged, since resonance effects are not taken into account. Secondly, a peculiar result is observed in the behavior of the “G” band when one uses a red light excitation. This Raman band exhibits a BreitWigner-Fano component on the low frequency side, modifying the band profile. This is due to a resonance effect since the energy of the laser excitation matches the electronic transition of metallic tubes, as demonstrated by Brown et al. [14]. The energy window for exciting metallic tube goes from 1.7 to 2.2 eV, while for other excitations, the Raman spectrum reveals mostly semiconducting tubes.

Figure 2. Calculations of the radial breathing mode frequency of (10,10) armchair single-walled nanotubes: a) isolated tubes; b) bundle of 7 tubes and c) bundle with an infinity of tubes. Carbon nanotubes therefore exhibit different electronic properties and their incorporation in electronic devices is rather promising. Nevertheless, it is of crucial importance to synthesize tubes in reproducible ways and to be

Raman and SERS Studies of Carbon Nanotubes


able to characterize them precisely. In this respect, and as far as Raman scattering is concerned, single-walled nanotubes are produced most of the time in bundles, and determining their diameter in a reliable manner is a difficult task. As a matter of fact, one has to take into account the Van der Waals interactions between tubes. Different models have been proposed, and in our case, we have described the carbon-carbon interaction by a Lennard Jones potential and we derived the tube-tube force constant from Cg, the adjacent inter-plane force constant of graphite. In these conditions, the diagonalization of the dynamical matrix leads to modifications of the radial breathing mode (RBM) frequencies, and the most intense one can be written as follows: 21 = 20 + C/mc , in which 0 is the frequency of an isolated tube, mc the mass of a carbon atom and a parameter which depends on the number of tubes constituting the bundle. Detailed calculations [13] have been published elsewhere and results can be viewed on Fig.2. We just recall here that for a bundle made of an infinite number of (10,10) armchair tubes, the upshift that is calculated is 16.2 cm-1. Even if such theoretical prediction is difficult to be proved experimentally, we have observed in isolated tubes prepared via a fluorination–defluorination procedure [15] that two RBM bands are observed at 177 and 194 cm-1, and the lowest in frequency (177 = 23 cm-1 is cm-1) is attributed to isolated tubes. The frequency shift rather close to our predictions. This result is corroborated by temperature effect measurements showing that these two components have a different behavior when the temperature is increased up to 600K [16].



SERS stands for Surface-Enhanced Raman Scattering. Introduced three decades ago [17] for studying monolayers, the SERS technique is used to provide a drastic amplification of the Raman signal. Most commonly, it is based on the use of a rough metallic surface that leads to such an amplification by resonantly exciting surface plasmons, the enhancement of the signal depending on the dielectric constant of the metal. Silver and gold turn out to be the most appropriate metals. It is known that the observed enhancement has a two-fold origin, one through an electromagnetic mechanism and the other having a chemical nature, i.e. being due to an increased polarisibility of the molecule adsorbed on the metal as a result of charge transfer processes. Depending on the method used for coupling to the surface plasmons, one can obtain amplification factors as high as 1012-1014 [18] although rough surfaces lead more commonly to 106-108. In our studies, we have used the SERS technique to amplify the Raman signal in an ultimate attempt to record Raman spectra on individual


S. Lefrant et al

nanotubes and prove experimentally the bundle effect, as explained above. In the case of SWNTs deposited on films of 100-150 nm in thickness, no significant differences are observed in comparison with Raman spectra of powders, if one excepts the amplification of the signal. On the contrary, when thinner films of nanotubes (below 100 nm) are used, SERS experiments have demonstrated the occurrence of interfacial reactions between nanotubes and the metallic support. The experiments have been carried out at different excitation wavelengths and the film thickness has been decreased from 150 nm down to 10 nm. Films were deposited on a

Figure 3. SERS spectra recorded at room temperature with exc.=676.4 nm of SWNTs deposited onto a rough gold substrate as a function of the film thickness: Curves 1, 2 and 3 are for 150nm, 60 nm and 30 nm, respectively.

gold or a silver substrate. In SERS spectra, the relative variations in intensity of some Raman lines, peak shifts and line shape changes resulting from the gradual decrease of the film thickness were studied [19], considered as originating from interactions between the substrate and the nanotube film. For laser wavelengths for which semiconducting tubes are predominantly excited (i.e. 514.5 or 1064 nm), the “D” band exhibits a relative increase of

Raman and SERS Studies of Carbon Nanotubes


intensity and the “G” band is not particularly affected neither in shape nor in frequency. This result can be interpreted by an increased disorder and/or formation on defects on nanotubes. Contrarily, using an excitation energy of 676.4 nm, for which the Raman response of metallic tubes is resonantly enhanced, a striking result is observed in the behaviour of the “D” and “G” bands, especially when Au is used as support. A gradual increase of the “D” band is recorded suggesting as before an enhanced degree of disorder, with a concomitant modification of the “G” band profile, consisting in a gradual narrowing of its low energy side (Fig.3). This modification can be interpreted again by a partial degradation of carbon nanotubes, leading to the formation to graphite-like or carbon particles. This result is further put in evidence if the film thickness decreases until 10 nm. We observe in the Raman spectrum the features attributed to amorphous carbon together with new bands characteristics of C60-like molecules. One may then consider the breaking of the nanotubes into species such as amorphous carbon, tubular fragments and closed-shell fullerenes. Such reactions are of chemical nature occurring at the nanotube-metal substrate interface. These results may appear as surprising. Nevertheless, previous studies reported in the literature show that SWNTs submitted to ball milling may be transformed into spherical carbon particles that are precursors of closed-shell fullerenes [20]. It has also been shown that electrochemical transformation of nanotubes can lead to the formation of C60 molecules [21]. These modification are indicative of a chemical mechanism in the SERS process in particular in very thin films for which chemical reactions are optimised. These experiments show in addition that metallic tubes are mainly involved, evidenced by the decrease of the Breit-Wigner-Fano component, as due to a possible direct interaction between nanotubes and C60 leading to the formation of SWNTs+C60- complexes [19].



Multi-walled nanotubes are made of concentric tubes whose number can be as high as 50. Prepared also by the electric arc technique, they can be purified in air at 600°C. High Resolution transmission Electron Microscopy images reveals clear images and in our case, we could for example statistically determine an average number of 18 tubes with inner diameters between 1 and 1.5 nm and outer diameters between 5 and 15 nm.


S. Lefrant et al

Such MWNTs exhibit Raman features (Fig. 4) that consist of the “G” band, the “D” band, but also low frequency modes whose origin is not straightforward.

Figure 4. Raman spectrum of multi-walled carbon nanotubes after purification recorded for exc. = 676.4 nm [22]. The low frequency modes have been calculated as explained in the text. In order to interpret these Raman characteristics, we have carried out calculations [23] by again introducing Van der Waals interaction between concentric tubes, on the basis on a Lennard-Jones potential as described in the case of bundles. It is obtained that modes that are seen in the low frequency domain originate from the RBM of individual tubes, the frequency of which being upshifted. In addition, we were able to evaluate the intensity of such modes by using a bond polarization theory [24]. We were able to reproduce Raman spectra taken at 647.1 and 514.5 nm by introducing different types of MWNTs. It is to be noticed that the main experimental parameters that have to be taken into account are the number of shells constituting the MWNTs and then the inner diameters. Such parameters have been extracted from HRTEM images. Our simulated Raman spectra [23] are

Raman and SERS Studies of Carbon Nanotubes


close to experimental ones, although a more sophisticated treatment would be needed to take into account the resonance effects that have been seen by different groups [25,26] when Raman spectra are recorded at different excitation wavelengths.



Polymer nanotubes composites are now extensively studied. Indeed, one may associate the properties of the polymer with those of nanotubes. This is the case of the mechanical reinforcement of standard polymer for example, but also one can take advantage of the specific electronic properties of the nanotubes. Therefore, we prepared composites with either saturated polymers like polymethylmethacrylate and MWNTs [27]. The electrical conductivity of these compounds as a function of the nanotube content exhibits for example a very low percolation threshold, (a few % in mass) and therefore they can be used as conducting and transparent layers in electronic devices such as Light Emitting Diodes (LEDs). Another type of composite that we have studied is based on the use of a conjugated polymer, polyphenylene-vinylene (PPV) known for its photoluminescence properties and SWNTs. We prepared this composite by mixing SWNTs to the precursor polymer of PPV. The conversion into PPV was subsequently performed by a thermal treatment at 300°C under dynamical vacuum [28]. We studied these materials by means of several optical spectroscopies including optical absorption, photoluminescence and Raman scattering in order to follow their properties as a function of the SWNT content. We observed drastic changes, as due to strong electronic interactions between SWNTs and the PPV precursor polymer. As an example, the photoluminescence spectra obtained at room temperature after excitation at 2.48 eV (500 nm) and 2.81 eV (440 nm) when the weight percentage of nanotubes x is varied from 0.5 to 64% are shown in Fig.5. When x is higher than 12, significant changes are observed, namely an increase of an emission peak at 2.43 eV, and the appearance of a new feature at 2.57 eV (482 nm) for x = 64 %. All the results have been interpreted by a model presented elsewhere [28] which takes into account the effective conjugation length distribution in the sample with the help of Raman spectra. Also, a blue-shift of the absorption band of the polymer in the composite is recorded. In brief, one can explain all these data by the formation of short conjugated segments whose relative weight dominate over longer segments, as due to the addition of SWNTs in the precursor polymer solution. In terms of applications, one can then monitor the maximum emission of the


S. Lefrant et al

compound by an appropriate amount of nanotubes. This technique is easier to achieve than modifying chemically the PPV polymer into a derivative whose emission properties are changed not as gradually as in the case of composites.

Figure 5. Luminescence spectra of poly-phenylene-vinylene/SWNTs composites as a function of the nanotube content expressed in mass percentage (x) from 0 to 64 %.



In this paper, we have shown the importance of Raman Scattering and Surface Enhanced Raman scattering to study carbon nanotubes together with High Resolution Transmission Electron Microscopy. This technique is a powerful tool to characterize different systems. In the case of SWNTs, one can deduce the diameter distribution in a specific sample by measuring the radial breathing mode frequency, and the nature of the tubes by changing the excitation wavelength. Using SERS, one can in addition test the model that uses Van der Waals interactions by allowing to record Raman spectra of individual tubes. Also, we could put in evidence interfacial reactions that

Raman and SERS Studies of Carbon Nanotubes


take place at the interface nanotube/metallic substrate, as a signature of a chemical component in the SERS mechanism. Raman scattering permits also to characterize multi-walled nanotubes and again, the introduction of Van de Waals interactions between the shells leads to a clear interpretation of the low frequency Raman modes. Finally, the study of different types of composites has been briefly discussed, as an example to exploit the properties of the carbon nanotubes to modify those of conducting polymers, and monitor their emission features as needed in devices like light emitting diodes. Acknowledgments: Samples of SWNTs used in our experiments were provided by Dr P. Bernier, University of Montpellier II. A large part of this work has been carried out in the frame of the European program COMELCAN (HRPN-CT-2000-00128) and a scientific collaboration between the Institut des Matériaux Jean Rouxel and the Laboratory of Optics and Spectroscopy of the National Institute of Bucharest. * The Institut des Matériaux Jean Rouxel is “Unité Mixte de Recherche” CNRS-Université de Nantes n° 6502 and is part of the “Ecole Polytechnique” of the University of Nantes.

1. Iijima S. , Nature (London), 354, 56 (1991). 2. Thess A. et al Science, 273, 483 (1996). 3. Journet C. et al., Nature, 388, 756 (1997). 4. See for example, Kong J., Cassell A.M. and Dai H., Chem.Phys.Lett. 292, 567 (1998). 5. See for instance "Physical properties of carbon nanotubes", ed. Saito R., Dresselhaus G., Dresselhaus M., Imperial College Press, London (1998). 6. Dresselhaus M.S., Jishi R.A., Dresselhaus G., Inomata D., Nakao K. and Saito R., Molecular Materials 4, 27 (1994). 7. Jishi R.A., Inomata D., Nakao K., Dresselhaus M.S. and dresselhaus G., J. Phys. Soc. Jpn 63, 2252 (1994). 8. Eklund P.C., Holden J.M. and Jishi A., Carbon 33, 959 (1995). 9. Rao A.M. , et al. Science, 275, 187 (1987). 10. Henrard L. et al., Phys. Rev. B 60, R8514 (1999). 11. Kahn et al., Phys. Rev. B 60, 6535 (1999). 12. Lefrant S., Buisson J.P., Chauvet O. and Benoit J.M., Proceedings of the MRS Fall meeting 2001, 706, Z 7.2 (2002). 13. Buisson J.P., Chauvet O ., Lefrant S., Stephan C. and Benoit J.M., Proceedings of the MRS Fall meeting 2000, 633, A.14.12.1 (2001).


S. Lefrant et al

14. Brown S.D.M., Corio P., Marucci A., Dresselhaus M.A., Pimenta M.A. and Kneipp K., Phys.Rev. 61, R5137 (2000). 15. Marcoux P., Schreiber J., Batail P., Lefrant S., Renouard J., Jacob G., Albertini D. and Mevellec J.Y., Phys. Chem. Chem. Phys. 4, 2278 (2002). 16. Schreiber J., Ph.D. thesis, University of Nantes, unpublished (2002). 17. Fleischmann M., Hendra P. and McQuillan A.J., Chem. Phys. Lett. 26, 163 (1974). 18. Kneipp K. et al., Phys. Rev. Lett. 84, 3470 (2000). 19. Lefrant S., Baltog I., Baibarac M., Schreiber J. and Chauvet O., Phys. Rev. B 65, 235401 (2002). 20. Dravid V.P., Lin X. ,Yang Y., WangX.K. ,Yee A. ,Ketterson J.B. and Chang R.P.H., Science,259, 1601,(1993). 21. Lefrant S. et al., to be published in Synth. Metals (2003). 22. Benoit J.M., Ph.D. thesis, University of Nantes, unpublished (2001). 23. Benoit J.M., Buisson J.P., Chauvet O., Godon C. and Lefrant S., Phys. Rev. B 66, 073417 (2003). 24. Guha S., Menendez J., Page J.B. and Adams G.B., Phys. Rev. B 53, 13106 (1996). 25. Jantoljak H., Salvetat J.P., Forro L. and Thomsen C., Appl. Phys. A: Mater. Sci. process 67, 113 (1998). 26. Kataura H., Achiba Y., Zhao X. and Ando Y., in Amorphous and Nanostructured Carbon, ed. By Robertson J. et al., Mater. Res. Soc. Symp. Proc. 593, 113 (2000). 27. Stephan C., Nguyen T.P., Lahr B., Blau W., Lefrant S. and Chauvet O., J. Mat. Res. 17, 396 (2002). 28. Wéry J., Aarab H., Lefrant S., Faulques E., Mulazzi E. and Pérego R., Phys. Rev. B 67, 115202 (2003).

Bing Zhao, Yukihiro Ozaki, The preparation and SERS study of triangular silver nanoparticles self-assembled films, J Raman Spectroscopy 2008, 39: 1673–...
近年来研 究的表面增强拉曼光谱(Surface Enhanced Raman Spectra,SERS)技术,通过群体 性地振荡金属自由电子, 使通过表面等离子体共振将空间自由传播的光局域在小 于...
(785nm)的 SERS,观察 G 峰的细节峰形结构来推测...Carbon Nanotubes and Graphene Raman Spectroscopy[J]...A Achete. Raman study of ion-induced defects in...
Crystal violet was used as probe molecule to study the relationship between ...(Surface enhanced Raman scattering, SERS)是一种特殊的表面 增强光学现象,通过...
08外文资料中文译文_能源/化工_工程科技_专业资料。不同形貌的银的制备及SERS...D) The Raman spectra from individual bare silver particles without MGITC ...
不同浓度的Cyt-C在W800上的Raman谱图 浦化工131009 冯...
[2] Xiaotian Wang, Wensheng Shi, Guangwei She and Lixuan Mua, Surface-Enhanced Raman Scattering (SERS) on transition metal and semiconductor nanostructures,...