Photoluminescence microscopy of carbon nanotubes grown by chemical vapor deposition: Influence of external dielectric screening on optical transition energies

Photoluminescence (PL) laser microscopy was applied to determine optical transition energies $E_{11}$ and $E_{22}$ of individual semiconducting single-walled carbon nanotubes (SWNTs) suspended on top of carbon nanotube “forests,” grown by chemical vapor deposition (CVD) on silicon substrates. A uniform increase of $E_{11}$ and $E_{22}$ energies by 40–55 and $24–48\mathrm{meV}$, respectively, was found for 19 different $(n,m)$ nanotube species suspended in air or a vacuum—relative to SWNTs in a reference water-surfactant dispersion. CVD-grown SWNTs embedded in paraffin oil and 1-methylnaphthalene show nearly the same PL peak positions as SWNTs in aqueous dispersion, indicating similar dielectric screening of excitons in SWNTs in these media.

Figure 1

SEM images of a $\sim 1\text{−}\mathrm{mm}$-high carbon nanotube forest grown by CVD on a silicon substrate at $750°\mathrm{C}$ for $90\min$. (see the text for growth details): (A) and (B) side views with different magnifications, (C) oblique view of the top surface. (D) High-resolution TEM image of typical nanotubes.

Figure 2

Histogram of diameters (heights above the surface) for individualized CVD-grown nanotubes spin coated onto a silicon surface from a sonicated aqueous dispersion and measured by AFM imaging (see the text for details).

Figure 3

(Color online) Typical microphotoluminescence maps (color-coded PL intensity vs excitation and emission wavelengths) measured on top of a $\sim 0.2\text{−}\mathrm{mm}$-high carbon nanotube forest. The maps were measured at different sample sites within three overlapping excitation ranges (see the Experimentals section). Some PL features are assigned to low- and high-energy sidebands $\left(S\right)$. Oblique stripes seen in the first two maps are Raman bands of carbon nanotubes.

Figure 4

(Color online) (a) Excitation-emission wavelength positions of $\sim 250$ peaks (filled squares) combined from microphotoluminescence maps of several samples of carbon nanotube forests measured in air and in a vacuum. The average positions are indicated by open circles. (b) Correlation of the PL peaks for SWNTs in air or a vacuum (open circles) to those observed in a water-surfactant dispersion and assigned to specific $(n,m)$ nanotubes (filled squares, Refs. 1, 24).

Figure 5

Shifts of the $E_{11}$ and $E_{22}$ energies for semiconducting SWNTs in air or a vacuum relative to aqueous dispersion($D_2\mathrm{O}∕1\mathrm{wt}\%$ SDBS) as a function of the nanotube diameter and the $(n-m)$ mod 3 value. Standard deviation bars show experimental uncertainty in determining the shifts (indicated for the shifts determined from at least five PL measurements; experimental error for PL peak positions of SWNTs in aqueous dispersion is neglected).

Figure 6

Comparison of PL peak positions for CVD-grown nanotubes immersed in paraffin oil and 1-methylnaphthalene (averaged from several micro-PL measurements) to SWNTs in aqueous dispersion (HiPco nanotubes in $D_2\mathrm{O}∕1\mathrm{wt}\%$ SDBS).