Determining single-walled carbon-nanotube redox states

Since the discovery of carbon nanotubes and their remarkable electronic, mechanical, and thermal properties, many groups have explored the origins of those properties and their applications in nanomaterials science and engineering.^1–3 Better understanding of the electronic properties of carbon nanotubes may facilitate development of better designs of electronic and optical devices. Single-walled carbon nanotubes (SWNTs) are made of a single layer of graphite rolled up into a tube. Chiral indices (n, m)describe the lattice points of the graphite sheet (quantifying how the tube is rolled up). The electronic structure of a SWNT strongly depends on its chiral indices.^1–6 However, to date it has not been possible to determine the electronic states of individual SWNTs with different chirality indices.

Figure 1. (top) 2D and (bottom) surface plots of photoluminescence (PL) intensity (in arbitrary units, a.u.) as a function of emission and excitation wavelength for a cast film of single-walled carbon nanotubes (SWNTs) on an indium tin oxide (ITO) electrode.

The oxidation/reduction potentials, bandgaps, Fermi levels, and work functions of SWNTs are critical quantities for understanding their electronic properties. We previously attempted to estimate the redox potentials of individual SWNTs by employing spectro-electrochemistry using visible-light/near-IR absorption and Raman spectroscopy. However, determination of the precise redox potentials was difficult because the SWNT spectral peaks overlapped. We have developed a new method and demonstrated our technique on 15 isolated SWNTs (see Figure 1) characterized by the chirality indices (n, m) = (6,5), (8,3), (7,5), (8,4), (10,2), (7,6), (9,4), (10,3), (8,6), (9,5), (12,1), (11,3), (8,7), (10,5), and (9,7).⁷ Crucially, in the near-IR regime, the photoluminescence (PL) peaks do not overlap, and so we were able to determine precise redox potentials of the individual (n, m)SWNTs.

We used in situ near-IR PL spectro-electrochemistry on films containing isolated (n, m)SWMTs that were prepared in an aqueous solution of carboxymethylcellulose and cast onto an indium tin oxide (ITO) electrode. We then added poly(diallyldimethylammonium chloride) to form an insoluble ion-complexed film. (It is necessary to have a film on the electrode that is insoluble in water when conducting spectro-electrochemistry in an aqueous system.)

We next applied an external potential in steps from 0.0V to −1.0V (reduction) and from 0.0V to +1.1V (oxidation). For both processes, we observed applied-potential-dependent PL quenching. The Nernst equation can be used to derive the redox potentials from our experimental data. Plots of normalized PL intensity against applied potential showed inflection points from which we can easily determine the oxidation and reduction potentials of the 15 isolated SWNTs. For instance, we plotted the normalized PL intensities of (7,5), (7,6), and (10,3)SWNTs as a function of external applied potential and the Nernst analysis curves (solid lines) of our experimental results (see Figure 2).

Figure 2. Normalized PL intensity of a film containing isolated SWNTs on an ITO electrode as a function of external applied potential and Nernst analysis curves (solid lines) of our experimental results. We applied a potential to the electrode in arbitrary steps from (a) 0.0V to −1.0V and (b) 0.0V to +1.1V. Ag∣AgCl: Standard silver/silver chloride solution electrode/electrolyte. e⁻: Electrons.

At the midpoints of the oxidation and reduction potentials, we can determine the Fermi levels of the 15 (n, m)SWNTs, and we used our results to find their work functions. We compared the latter, determined based on first-principles density functional theory, with our experimentally determined values. The results correlated well. In particular, we found that the work functions of SWNTs with diameters greater than 0.85nm changed significantly with varying chirality. In addition, for SWNTs with diameters smaller than 0.85nm, the work functions increased with decreasing diameter.

We emphasize that our method is very simple and can be used to determine precise electronic states of any isolated SWNT with detectable PL. We are now doing just this for many different kinds of SWNT. It is also possible to control the desired electronic states of isolated SWNTs by fine tuning the external, applied potential, which may help improve the design and fabrication of electronic nanodevices using SWNTs.

We are grateful to Susumu Saito (Tokyo Institute of Technology, Japan) for his collaboration in calculating work functions.