Charge transfer and Fermi level shift in $p$-doped single-walled carbon nanotubes

The electronic properties of $p$-doped single-walled carbon nanotube (SWNT) bulk samples were studied by temperature-dependent resistivity and thermopower, optical reflectivity, and Raman spectroscopy. These all give consistent results for the Fermi level downshift $\left(\Delta E_F\right)$ induced by doping. We find $\Delta E_F\approx 0.35\mathrm{eV}$ and $0.50\mathrm{eV}$ for concentrated nitric and sulfuric acid doping respectively. With these values, the evolution of Raman spectra can be explained by variations in the resonance condition as $E_F$ moves down into the valence band. Furthermore, we find no evidence for diameter-selective doping, nor any distinction between doping responses of metallic and semiconducting tubes.

Figure 1

The x-ray diffraction data from (a) PLV and (b) HiPco buckypapers. The PLV material is more crystalline than HiPco and has a narrower diameter distribution. The major effect of doping is the loss of Bragg intensity, more dramatic for sulfuric than nitric acid. No difference in doping response was found between PLV and HiPco, thus no evidence for diameter selectivity.

Figure 2

(Color online) The doping effect on resistivity and its temperature dependence. Doping reduces $\rho \left(300\mathrm{K}\right)$, approximately twice as much for sulfuric than nitric acid. The temperature dependence becomes weaker after doping, more dramatically so for sulfuric than nitric acid.

Figure 3

(Color online) The doping effect on thermopower. For undoped and air-exposed SWNT S(T) increases monotonically to $60–80\mu \mathrm{V}∕\mathrm{K}$ at $300\mathrm{K}$ (not shown). Doping reduces the amplitude at all $T$ and introduces an inflection point at characteristic temperatures in the range $20–50\mathrm{K}$ for different samples. Both effects are most apparent in the ${\mathrm{H}}_2\mathrm{S}{\mathrm{O}}_4$-doped HiPco fiber.

Figure 4

(Color online) The reflectivity and Drude-Lorentz fits for undoped, $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$-doped, and ${\mathrm{H}}_2\mathrm{S}{\mathrm{O}}_4$-doped PLV (a) and HiPco (b) buckypapers. Two systematic effects are observed: a blueshift of the reflectivity minimum (increasing Drude plasma frequency) and decreased intensity of the peaks at higher energy (quenching interband transitions by shifting $E_F$). See Table for model fit parameters (solid curves).

Figure 5

The Raman spectra of undoped, $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$-doped, and ${\mathrm{H}}_2\mathrm{S}{\mathrm{O}}_4$-doped PLV buckypapers using (a) $2.41\mathrm{eV}$ and (b) $1.96\mathrm{eV}$ laser excitation. The RBM bands are magnified for clarity. Raman intensities decrease upon doping, more dramatically for sulfuric acid.

Figure 6

The Raman spectra of HiPco fibers using (a) $2.41\mathrm{eV}$, (b) $1.96\mathrm{eV}$, and (c) $1.59\mathrm{eV}$ laser excitation. The RBM bands are magnified for clarity. “Low” and “high” concentrations of ${\mathrm{H}}_2\mathrm{S}{\mathrm{O}}_4$ were obtained using different coagulation baths in the spinning process. The loss of resonance upon doping is similar to PLV (Fig. 5). See Table for estimated $E_F$ shifts associated with loss of Raman resonance.