Alkali-metal-doping dynamics and anomalous lattice contraction of individual debundled carbon nanotubes

The dynamics of K doping of thin films of semiconducting debundled single-walled carbon nanotubes (SWNTs) on a $\mathrm{Si}∕{\mathrm{SiO}}_2$ substrate is followed using in situ Raman scattering. Similar to the behavior we observed in the Cs doping of bundled SWNTs, four distinct intervals during doping can be distinguished via the behavior of the high frequency tangential band (or G-band) scattering. We therefore find that the character of the time evolution of the frequency and linewidth of the most prominent G band with alkali-metal doping is almost identical for debundled and bundled tubes. This includes the observation of an anomalous upturn in the G-band frequency during the third interval of doping that is consistent with a contraction of the $\mathrm{C}\mathrm{C}$ bond within the tube wall. We conclude that this anomalous contraction is not associated with doping-induced changes in the tube-tube interaction within a bundle, but is an intrinsic, and yet unexplained, consequence of an upshifting Fermi level in debundled $n$-doped SWNTs. A long induction period between when doping conditions are established and when changes in the Raman spectrum occur is observed and remains unexplained.

Figure 1

(Color online) Typical AFM image of the sample comprised of $\sim 50\%$ individual SWNTs. The empty circles represent the nanotube morphology according to the AFM analysis, i.e., whether SWNTs exist as individual tubes or as small bundles containing two or three individual tubes.

Figure 2

(Color online) Time evolution of the Raman spectra with in situ K doping (Interval I) of debundled SWNTs. The R-band spectra have been expanded and placed on top of each full spectrum for better observation. The inset shows the schematic drawing of the experimental apparatus. The length of the Pyrex reactor (inset) is $\sim 25\mathrm{cm}$. 514.5 nm (2.41 eV) laser excitation was used for the Raman study.

Figure 3

Time evolution of the Raman spectra with in situ K doping (Interval II) of debundled SWNTs. The R-band spectra have been expanded and placed on top of each full spectrum for better observation.

Figure 4

Time evolution of the Raman spectra with in situ K doping (Interval III) of debundled SWNTs. The dots are the experimental data, the thicker solid lines are results of a Lorentzian line-shape analysis to the data, and the peaks underneath are individual Lorentzian components used for the fit.

Figure 5

Time evolution of the Raman spectra during Interval IV of (a) K doping of debundled SWNTs, and (b) Cs doping of bundled SWNTs. The dots are the experimental data, and the thick solid lines are the results of either a combination of Lorentzian and Breit-Wigner-Fano (BWF) line-shape analysis (left panel) or a single BWF line fit (right panel). The individual peaks, i.e., a Lorentzian component (solid line) and a BWF peak (dashed line), are shown underneath the overall fit in the left panel. The peak marked with Hg in the right panel is from the Hg lamp used for frequency calibration.

Figure 6

Time evolution of the fitting parameters ${\omega }_T$ and ${\Gamma }_{\mathrm{T}}$ used to describe the dominant G band. The top is for Cs doping of bundled SWNTs (Ref. 7) and the bottom is for K doping of debundled SWNTs. The solid squares represent fitted data of ${\omega }_T$, and the stars are fitted data of ${\Gamma }_T$. Polynomial fitted curves (solid line) were used as a guide to the eye (see text). In Intervals I–III, ${\omega }_T$ and ${\Gamma }_{\mathrm{T}}$ are the center frequency and peak width. For Interval IV, ${\omega }_T$ and ${\Gamma }_{\mathrm{T}}$ are the renormalized phonon frequency and the Breit-Wigner-Fano line broadening parameter [Eq. (1)].