Identification of an excitonic phonon sideband by photoluminescence spectroscopy of single-walled carbon-13 nanotubes

We have studied photoluminescence and resonant Raman scattering of single-walled carbon nanotubes (SWNTs) consisting of carbon-13 $\left({\mathrm{SW}}^{13}\mathrm{CNTs}\right)$ synthesized from a small amount of isotopically modified ethanol. There was almost no change in the Raman spectra shape for ${\mathrm{SW}}^{13}\mathrm{CNTs}$ except for a downshift of the Raman shift frequency by the square root of the mass ratio $12∕13$. By comparing photoluminescence excitation spectra of ${\mathrm{SW}}^{13}\mathrm{CNTs}$ and normal SWNTs, the excitonic phonon sideband due to strong exciton-phonon interaction was clearly identified with the expected isotope shift. In addition to the direct experimental proof of the strong exciton-phonon interaction, we also found low-intensity “pure electronic” features whose origin has never been elucidated.

Figure 1

(Color online) Raman spectra measured with a $488\mathrm{nm}$ $\left(2.54\mathrm{eV}\right)$ excitation laser and PL maps of normal SWNTs and ${\mathrm{SW}}^{13}\mathrm{CNTs}$. (a) RBM, (b) $G$ band. Dotted lines are spectra of normal SWNTs shifted by multiplying the Raman shift frequency by the mass ratio factor $\sqrt{12∕13}$. PL maps of (c) normal SWNTs and (d) ${\mathrm{SW}}^{13}\mathrm{CNTS}$.

Figure 2

(Color online) PL maps of (a) normal SWNTs and (b) ${\mathrm{SW}}^{13}\mathrm{CNTs}$ dispersed in surfactant suspension. (c) Comparison of PLE spectra of ${\mathrm{SW}}^{13}\mathrm{CNTs}$ and normal SWNTs at the $E_{11}$ emission energy of (7, 5) nanotubes. Each PLE spectrum was measured along the PL emission energy of $1.208\mathrm{eV}$.

Figure 3

(Color online) Comparison of (a) PL emission spectra at the $E_{22}$ transition energy of (7, 5) nanotubes $\left(1.923\mathrm{eV}\right)$, (b) the PLE spectra of (7, 5) nanotubes around $E_{22}$ transition energy corresponding to the different emission energies, and (c) magnifications of the PLE spectra around peak $C$ for different PL emission energies plotted as a function of energy distance from the $E_{22}$ peak maxima. Each spectrum intensity is normalized by $E_{22}$ intensity and the peak-top intensities are leveled for comparison in (c). Vertical dashed lines and horizontal two-headed arrows in (a) indicate the central energies of the detection slit and PL emission slit width. The red circle and black cross correspond to ${\mathrm{SW}}^{13}\mathrm{CNTs}$ and normal SWNTs, respectively. Solid (dotted) bars in (c) represent the phonon energies of the LO phonons near the $K$ $\left(\Gamma \right)$ point of the graphene Brillouin zone (Refs. 18, 25). The colors of the bars correspond to normal SWNTs (black) and ${\mathrm{SW}}^{13}\mathrm{CNTs}$ (red).

Figure 4

(Color online) PLE spectra for peak $A$ of (7, 5) nanotubes plotted as a function of energy difference from the $E_{11}$ energy. Vertical line around $\sim 0.2\mathrm{eV}$ indicates the energy corresponding to Raman scattering ($G$ band).

Figure 5

(Color online) (a) PLE spectra of (6, 5) nanotubes around $E_{22}$ transition energy, and (b) magnifications of the PLE spectra around peak $C^{\prime }$ for (6, 5) nanotubes plotted as a function of energy difference from the $E_{22}$ energy. Each spectrum-intensity is normalized by $E_{22}$ intensity and the peak-top intensities are leveled for comparison in (b).