Anharmonicity in single-wall carbon nanotubes as evidenced by means of extended energy loss fine structure spectroscopy analysis

A comparative study of the structure of free-standing parallel bundles of single-wall carbon nanotubes (SWCNTs), multiwalled carbon nanotubes (MWCNTs), and highly oriented pyrolytic graphite (HOPG) was achieved by means of transmission electron microscopy and electron energy loss spectroscopy analyses. In particular, the carbon $K$ $\left(1s\right)$ extended fine structure of SWCNTs is found to be characterized by an apparent contraction of the nearest neighbors distance. This contraction is interpreted here to originate from an asymmetric pair distribution function, mostly due to the high out-of-plane vibrational motion of the C atoms, as for the case of chemisorbed atoms on clean surfaces. In contrast, the MWCNTs did not exhibit any signature of such an anharmonic effect because of their more rigid structure. This indicates that the SWCNTs pair potential is significantly broader and its effect is much weaker than that experienced by the same $\mathrm{C}\text{−}\mathrm{C}$ pair embedded in a multiwall nanotube.

Figure 1

Transmission electron micrographs of a $12\mathrm{nm}$-diam. MWCNT (a) and a $(22\pm 2)\mathrm{nm}$-diam. bundle of SWCNTs, of about $1.2\mathrm{nm}$-diam. each (Ref. 20) (b). Since all the walls appear to be parallel and straight, the presence of structural defects (such as pentagons and heptagons) in the carbon hexagonal network can be considered to be negligible. In the inset we report a higher magnification of the TEM image of the inner walls of the MWCNT, showing a distance of about $0.37\mathrm{nm}$ between each graphene sheet.

Figure 2

(Color online) Electron energy loss spectra, after background subtraction, at the carbon $K$ edges for HOPG, the MWCNT, and the SWCNTs rope. No nitrogen or oxygen $K$ edge related features (at 409 and $530\mathrm{eV}$, respectively) can be detected in all the three spectra.

Figure 3

(Color online) (a) EXELFS signals, $k\chi \left(k\right)$, and (b) their corresponding Fourier transforms, $F\left(R\right)$, for HOPG, a SWCNTs bundle, and a MWCNT. The $k$ range for the $F\left(R\right)$ curve calculations is extended from approximately $k_{\min }=4.4{\mathrm{\AA }}^{-1}$ to $k_{\max }=11{\mathrm{\AA }}^{-1}$.

Figure 4

(Color online) EXELFS signal $k{\chi }_1\left(k\right)$ only from nearest-neighbor shell after Fourier filtering (approximately from $R_{\min }=0.5\mathrm{\AA }$ to $R_{\max }=1.7\mathrm{\AA }$) for HOPG (solid line), a SWCNTs bundle (dot-dashed line), and a MWCNT (dashed line).

Figure 5

(a) Phase difference of the EXELFS signal $k{\chi }_1\left(k\right)$, only from nearest-neighbor shell, between HOPG and both types of nanotubes [MWCNTs (circles) and SWCNTs rope (triangles), respectively]. (b) Logarithmic ratio of the backscattering amplitude of SWCNTs rope versus HOPG (triangles) and MWCNTs versus HOPG (circles). The solid curves represent the best fit to the experimental data.