wavenumber (cm- )

Figure 11. Raman spectra of SWNTs grown at increasing reaction times. The product was obtained at 750 °C over a CoMo catalyst by CO disproportionation. Reprinted with permission from [32], W. E. Alvarez et al., Chem. Mater. 14, 1853 (2002). © 2002, American Chemical Society.

[85, 104, 120-123]. This tangential mode G band appears in the 1400-1700 cm-1 region and involves out-of-phase intralayer displacement in the graphene structure of the nanotubes. In contrast to the D band the G band is a measure of the presence of ordered carbon and therefore is present on all kind of sp2 ordered carbonaceous materials (graphite, MWNTs, nanofibers, etc).

In SWNT samples, the most important characteristic of the G band is that the analysis of its line shape offers a method for distinguishing between metallic and semiconducting nanotubes [86, 88, 108]. The G band of the semiconducting nanotubes has been extensively studied and is well accounted for using Lorentzian oscillators to describe the six Raman active modes that have been spectroscopi-cally identified by polarization studies of the symmetries of the various line-shape components [124, 125]. On the other hand, although Lorentzians can be used to describe the G band of metallic carbon nanotubes [75, 126, 127], the lower frequency component of the G band spectrum that appears around 1540 cm-1 has to be fitted using a Breit-Wigner-Fano (BWF) line shape, due to a downshift and broadening in the tangential G band of metallic SWNTs relative to semiconducting SWNTs [78, 128-130].

Brown et al. [131] demonstrated that only these two Raman components were needed to fit the tangential G band for metallic SWNTs. Both components were found to exhibit predominant A(A1g) symmetry. The differences in their peak frequencies were attributed to a difference in the force constant of vibrations along the nanotube axis (higher force constant) versus circumferential (lower force constant) and to an additional downshifting and broadening of the lower frequency peak due to a coupling of the discrete phonons to an electronic continuum (this continuum is characteristic of any electronic structure that shows metallic-like conduction properties) that resulted in the BWF line shape. However, as in all resonant Raman studies, the presence of this BWF in the Raman spectra will depend on the resonance conditions between the excitation energy and the

DOS of the SWNT present on the sample. Accordingly, in a sample in which SWNTs of different chirality are present it is possible to selectively enhance the Raman efficiency for metallic or semiconductor SWNTs by carefully choosing the laser excitation energy. This method will be illustrated.

Figure 12 shows a wide range resonant Raman spectra of the G band for SWNTs obtained by arc discharge using a RhPd catalyst [78]. For laser excitations around 2.4 eV the G band takes the shape of a broad and asymmetric peak centered at 1540 cm-1 with a BWF line shape. This band shape undoubtedly indicates that metallic nanotubes are in resonance with the laser excitation energy. In this case, the SWNT in the sample had diameters between 0.7 and 1.0 nm, as observed by TEM and inferred from the A1g breathing mode frequencies of the Raman spectra. Consequently, when using laser energies around 1.8 eV, mostly metallic SWNTs are probed and a "metallic window" can be delineated in which the resonance condition for metallic nanotubes with diameter around 0.9 nm takes place. This metallic window is depicted in the Kataura plot of Figure 8.

Pimenta et al. [75] have investigated the line-shape variation of the band attributed to metallic nanotubes as a function of the excitation laser energy. In this study, a large enhancement in the relative intensity of the band centered at 1540 cm-1 was observed in the energy range 1.72.2 eV This enhancement is shown in Figure 13, and it was explained in terms of the singularities in the density of states of the metallic nanotubes together with the distribution of

Figure 12. Resonance Raman spectra of SWNTs synthesized using RhPd by arc discharge. The diameter distribution ranges from 0.7 to 1.00 nm. Reprinted with permission from [78], M. Kataura et al., Synth. Met. 103, 2555 (1999). © 1999, Elsevier Science.

1400 1500 1600 1700

Raman shift (cm'1)

Figure 13. Raman spectra of the tangential modes of SWNTs obtained with several laser energies. The SWNTs were synthesized by laser ablation of a carbon target containing a Ni/Co catalyst. Reprinted with permission from [75], M. A Pimenta et al., Phys. Rev. B 58, R16016 (1998). © 1998, American Physical Society.

1400 1500 1600 1700

Raman shift (cm'1)

Figure 13. Raman spectra of the tangential modes of SWNTs obtained with several laser energies. The SWNTs were synthesized by laser ablation of a carbon target containing a Ni/Co catalyst. Reprinted with permission from [75], M. A Pimenta et al., Phys. Rev. B 58, R16016 (1998). © 1998, American Physical Society.

diameters of the sample. The nanotubes used in that particular study were obtained by laser vaporization of a carbon target containing 1 to 2% of a Ni/Co catalyst. The diameter distribution obtained from the TEM was in the 1.1-1.3 nm range and, as predicted from the plot in Figure 8, for those diameters the energy region of resonance for metallic nano-tubes falls in the range 1.7-2.2 eV. This energy range is in perfect agreement with the range at which the enhancement of the BWF line shape was observed.

As a third example, Figure 14 shows the set of Raman spectra on the G band for nanotubes produced by catalytic CO disproportionation at 950 °C on Co-Mo catalysts. Under these conditions, the SWNTs produced have an average diameter of 1.7 nm [32, 88]. Consequently, an enhanced BWF line was observed at higher laser excitation energies (2.4-2.55 eV) than for the previous cases. When a laser with lower excitation energy (e.g., 2 eV) was used, the broad band at 1540 cm-1 greatly decreased its relative contribution. This decrease can be explained taking into account that, in this case, one moves out of the metallic window (see Fig. 8).

In spite of all the interesting information that can be obtained from Raman spectroscopy, this technique has an intrinsic limitation related to its resonant character, which permits one to only get information on the nanotubes that are in resonance with the incident or scattered light. Therefore, a full characterization of all the SWNTs in the sample, with different diameters and chiralities, cannot be achieved.

Only on those samples with a narrow diameter distribution or, even better, on isolated SWNTs can Raman spectroscopy provide a full description of the sample. Recently, a series of studies on isolated SWNTs grown over a silicon substrate has been reported. These stud ro >

1400 1500 1600 1700


Figure 14. Resonant Raman spectra in the tangential mode (G mode) obtained using three different laser excitation energies, for SWNT samples produced by CO disproportionation at 950 °C. Reprinted with permission from [88], J. E. Herrera et al., J. Nanotech., in press. © American Scientific Publishers.

ies have led to the observation of Raman spectra from an isolated nanotube, with intensities under good resonance conditions comparable to those from the silicon substrate, even though the ratio of carbon to silicon atoms in the light beam was approximately only 1 carbon atom to 108 silicon atoms [132-137]. All the Raman features observed in previous studies on nanotube bundles were also observed in the spectra of the single nanotubes, including the radial breathing mode, the G band, and the D band [132, 133]. However, at the single nanotube level, the characteristics of each feature were investigated in greater detail, including its dependence on diameter, chi-rality, laser excitation energy, and closeness to resonance with electronic transitions [134]. In this case, the uniqueness of the electronic transition energies for each nanotube becomes of particular importance [135]. The high sensitivity of the Raman spectra to diameter and chirality, particularly for the characteristics of the radial breathing mode, which are also uniquely related to the (n, m) indices, provided structural determination of (n, m) at this single nanotube level. The (n, m) assignments made to individual carbon nanotubes have been corroborated by measuring the characteristics of other features in the Raman spectra that are also sensitive to nanotube diameter and chirality. Raman spectroscopy has provided a convenient way to characterize nanotubes for their (n, m) indices, in a manner that is compatible with the measurement of other nanotube properties, such as transport, mechanical, and electronic properties at the single nanotube level, and the dependence of these properties on nanotube diameter and chirality [136, 137].

0 0

Post a comment