Resonance Raman Scattering of MWNTs

Multi-walled carbon nanotubes (MWNTs) prepared by the carbon arc method are thought to be composed of a coaxial arrangement of concentric nanotubes. For example, 13C-NMR [64] and magnetoresistance measurements [65,66] show Aharonov-Bohm effects that are associated with the concentric tube structures. On the other hand, the thermal expansion measurements [67] and the doping effects [68] suggest that some kinds of MWNTs have scroll structures. If the RBMs, which are characteristic of SWNTs [35], are observed in MWNTs, the RBM Raman spectra might provide experimental evidence for the coaxial structure. In many cases, however, MWNTs have very large diameters compared with SWNTs even for the innermost layer of the nanotube, and no one has yet succeeded in observing the RBMs in large diameter MWNTs. Zhao and Ando have succeeded in synthesizing MWNTs with an innermost layer having a diameter less than 1.0 nm, by using an electric arc operating in hydrogen gas [69]. The spectroscopic observations on this sample revealed many Raman peaks in the low frequency region, which these RBM frequencies can be used to assign (n, m) values for some constituent layers of MWNTs [70]. Since the resonance Raman effect can be observed in MWNTs (see Fig. 20), we can be confident that these low frequency features are associated with RBMs.

Several MWNT samples have been prepared by the carbon arc method using a range of hydrogen pressures from 30 to 120Torr, and yielding good MWNT samples under all of these operating conditions. Relative yields depend on the hydrogen pressure and on the arc current [69], with the highest

1111111111 11 11 11 11 11 111111ii|i 11111

Excitation (eV) 2.73

1111111111 11 11 11 11 11 111111ii|i 11111

Excitation (eV) 2.73

Fig. 20. The resonant Raman spectra of multi-wall carbon nanotubes with very small innermost diameters that grow preferentially using an electric arc in hydrogen gas [71]

1 50 200 250 300 350 400 450 500 550

1 50 200 250 300 350 400 450 500 550

Fig. 20. The resonant Raman spectra of multi-wall carbon nanotubes with very small innermost diameters that grow preferentially using an electric arc in hydrogen gas [71]

yield of MWNTs being obtained at 60Torr of hydrogen gas pressure. The sample purity, after purification of the sample, which was characterized using an infrared lamp, is over 90% MWNTs and the diameter distribution of the innermost shell was measured by TEM. Most of the MWNTs have diameters of the innermost shell of about 1.0 nm, and sometimes innermost diameters less than 0.7 nm were observed.

In Fig. 20 resonance Raman scattering of samples synthesized under different conditions have been measured, and RBM peaks have been observed from 200 to 500 cm-1 [63]. Peaks between 150 and 200 cm-1 are due to the air. In fact these peaks of O2 and N2 are commonly observed not only for the MWNT sample but also for the quartz substrate and they are not observed in Ar gas. Peaks above 200 cm-1 show very sharp resonances, which strongly suggest that these structures originate from the RBM vibrations of nano-tubes. Resonance effects for each peak are similar to those of single-walled nanotubes. However, the peak frequencies are about 5% higher than those of single-walled nanotubes with the same diameter, which might be due to the inter-layer interaction. For example, the RBM peak at 280 cm-1 shows a maximum intensity at 2.41 eV. This is the same behavior as the peak at 268 cm-1 in SWNTs. This fact is consistent with the recent calculation of bundle effects on the RBM frequency of SWNTs which predict a 10% upshift in the mode frequency due to tube-tube interactions [53]. From the simple relationship between nanotube diameter and RBM frequency [35], the candidate nanotube shells for the peak at 490cm-1 are (5,1), (6,0), (4,3) and (5,2) having RBM frequencies at 509.6, 472.8, 466.4 and 454.3cm-1, respectively. If we take into account the 5% up-shift due to the interlayer interactions, the candidates are narrowed down to the nanotube shells (6,0), (4,3) and (5,2), which have diameters of 0.470, 0.477 and 0.489 nm, respectively, and these diameters are consistent with the TEM observations. It is very interesting that the (6,0) nanotube has the same structure as D6h C36 which has D6h symmetry [72]. However, we also have to consider the electronic states of the nanotube to clearly identify the resonance effect. By use of the zone-folding band calculation [2,8,16], assuming a transfer integral 70 = 2.75eV, it is found that (6,0) and (5,2) are metallic nanotubes and have their lowest energy gap E^ at 4.0 eV. The resonance laser energy, where the RBM peak has a maximum intensity, occurs at 1.7eV, and the peak at 490 cm-1 was assigned to the (4,3) nanotube which is a semiconductor, and has its lowest energy gap ES1 at 1.6eV. In the same way, the candidates (7,1) and (5,4) were considered for the Raman band at 388 cm-1. The nanotube (7,1) is metallic and the lowest energy gap E^J is at 3.4 eV, while the (5,4) nanotube should be semiconducting and is expected to have Ef1 and Ef2 at 1.28 and 2.52 eV, respectively. Thus, the peak at 388 cm-1 should be assigned to the nanotube shell (5,4) because of the resonance observed at 2.4eV[71].

Finally we consider the interlayer interactions in MWNTs. The RBM band in Fig. 20 at 490 cm-1 is split into three peaks indicating the same resonance feature. These peaks cannot be explained by different nanotubes, since there are no other candidates available. The nanotube (5,1) is the only candidate having the nearest diameter and the nearest energy gap in the optical spectra. However, the calculated energy gap of a (5,1) nanotube is 1.7 eV, which is 0.1 eV wider than that for a (4,3) nanotube. If one of the peaks originates from a (5,1) nanotube, the resonance feature should be different from that for the other peaks. Further, the RBM frequency of a (5,1) nanotube becomes 534 cm-1, taking into account the 5% up-shift due to the inter-tube interaction in a nanotube bundle. Thus, it is proper to think that these three peaks are originating from the same nanotube. The possible reason for the splitting of this peak is the interlayer interaction. When the first layer is (4,3), then (10,7) is the best selection as the second layer, since the interlayer distance is 0.342 nm, which is a typical value for MWNTs [73]. The other nearest candidates for the second layers are (13,3), (9,8) and (11,6) having inter-layer distances 0.339, 0.339 and 0.347nm, respectively. The interlayer distance for the (13,3) and (9,8) candidates are about the same (about 1% smaller) as the typical inter-layer distance, and but the interlayer distance for the (11,6) nanotube is 1.5% larger. The magnitude of the interlayer interaction should depend on the interlayer distance, and, consequently, the RBM frequency of the first layer may depend on the chiral index of the second layer. Indeed, the observed frequency separation between the split peaks is about 2%, which may be consistent with the difference in interlayer distances. The splitting into three RBM probably indicates that there are at least three kinds of second layers. Furthermore, this splitting cannot be explained by a scrolled structure for MWNTs. This strongly suggests that the MWNTs fabricated by the electric arc operating in hydrogen gas has a concentric structure. For the thinnest nanotube (4,3), the RBM frequency of the second layer is 191 cm-1. This should be the highest RBM frequency of the second layer nanotube. Since the low frequency region is affected by signals from the air, Raman spectra were taken while keeping the sample in argon gas. However, no peak was observed below 200 cm-1, suggesting that only the innermost nanotube has a significant Raman intensity. The innermost layer has only an outer nanotube as a neighbor, while the other nanotubes, except for the outermost layer, have both inner and outer nanotube neighbors.

The interlayer interaction probably broadens the one-dimensional band structure, in a like manner to the bundle effect in SWNTs [48,49,50,51,52,29] [53,54]. The band broadening decreases the magnitude of the joint density of states at the energy gap, leading to a decrease in the resonant Raman intensity of the second layer. On the other hand, the RBM frequency of the outermost layer is too low to measure because of its large diameter. Thus, RBMs are observed in MWNTs only for the innermost nanotubes. Theoretical calculations show that SWNTs with diameters smaller than Ceo show metallic behavior because of the hybridization effect of the 2pz orbital with that of the a electron [74,75]. The hybridization effect lowers the energy of the conduction band and raises the energy of the valence band, which results in the semi-metallic nature of the electronic states. However, the electrostatic-conductance of two-probe measurement of MWNTs shows that semiconducting nanotubes seems to be dominant in this diameter region[76]. Thus it is necessary to investigate the electronic properties of SWNTs with diameters smaller than that of C60.

0 0

Post a comment