Selective chemistry of single-walled nanotubes

The first extensive SWNT sidewall reaction was fluorination in 1998 by Mickleson et al. These fluorine moieties can be removed from the nanotube by treatment in hydrazine and the spectroscopic properties of the SWNT can be restored completely.[2]

One of the most important SWNT sidewall reaction is that with diazonium reagent which if done under controlled conditions can be used to do selective covalent chemistry.

Water-soluble diazonium salts react with carbon nanotubes via charge transfer in which they extract electrons from SWNT and form a stable covalent aryl bond. This covalent aryl bond forms with extremely high affinity for electrons with energies near the Fermi level, E_f of the nanotube. Metallic SWNT have a greater electron density near E_f resulting in their higher reactivity over semiconducting nanotubes. The reactant forms a charge-transfer complex at the nanotube surface, where electron donation from the latter stabilizes the transition state and accelerates the forward rate. Once the bond symmetry of the nanotube is disrupted by the formation of this defect, adjacent carbons increase in reactivity and initial selectivity for metallic SWNT is amplified. Under carefully controlled conditions this behavior can be exploited to obtain highly selective functionalization of metallic nanotubes to the near exclusion of the semiconductors.[3][4]

Primary condition is addition of reactant molecules at a very small rate to SWNT solution for a sufficient long time. This ensures reaction with only metallic SWNTs and with no semiconducting SWNTs as all the reactant molecules are taken up by the metallic SWNTs. Long time injection ensures that all metallic tubes are reacted. For example, one highly selective condition is: addition of 500 µL of 4-hydroxybenzene diazonium tetrafluoroborate solution in water (0.245 mM) at an injection rate of 20.83 µL/h into 5 mL of SWNT solution (1 wt % sodium dodecyl sulfate (SDS)) over 24 hrs. However, if the entire diazonium solution is added all instantaneously then semiconducting SWNTs will also react due to presence of excess reactant.[4]

Optical absorption monitors the valence (v) to conduction (c) electronic transitions denoted E_nn where n is the band index. The E₁₁ transitions for the metallic nanotubes occur from ~440 to 645 nm. The E₁₁ and E₂₂ transitions for the semiconducting nanotubes are found from 830 to 1600 nm and 600 to 800 nm, respectively. These separated absorption features allow for the monitoring of valence electrons in each distinct nanotube. Reaction at the surface result in localization of valence electrons makes them no longer free to participate in photoabsorption which results in decay of the spectrum features.[3][5]

Selective diazonium chemistry abruptly decreases the peak intensities that represent the first Van Hove transition of metallic species (E₁₁, metal), while the peak intensities representing the second (E₂₂, semiconducting) and first (E₁₁, semiconducting) Van Hove transition of the semiconducting species show little or no change. A relative decrease in metallic SWNT absorption features over semiconducting features represents a highly preferential functionalization of the metallic nanotubes.[3][5]

Raman spectroscopy is a powerful technique with wide-ranging applications in carbon nanotube studies. Some important Raman features are radial breathing mode (RBM), tangential mode (G-band), and disorder-related mode.

RBM features correspond to the coherent vibration of the C atoms in the radial direction of the nanotube. These features are unique in carbon nanotubes and occur with the frequencies ω_RBM between 120 and 350 cm⁻¹ for SWNT in the diameter range (0.7 nm-2 nm). They can be used to probe the SWNT diameter, electronic structure through their frequency and intensity (I_RBM) respectively and hence perform an (n,m) assignment to their peaks. The addition of the moiety to the sidewall of the nanotube disrupts the oscillator strength that gives rise to RBM feature and hence causes decay of these features. These features are distinct for species of a particular nanotube (n,m) and hence enables to probe which SWNTs are functionalized and to what extent.

Two main components of the tangential mode include G⁺ at 1590 cm⁻¹ and G⁻ at 1570 cm⁻¹. G⁺ feature is associated with carbon atom vibrations along the nanotube axis. The G⁻ feature is associated with vibrations of carbon atoms along the circumferential direction. The G-band frequency can be used (1) to distinguish between metallic and semiconducting SWNTs, and (2) to probe charge transfer arising from doping a SWNT. Frequency of G⁺ is sensitive to charge transfer. It upshifts for acceptors and downshifts for donors. Lineshape of G⁻ is highly sensitive to whether SWNT is metallic (Breit-Wigner-Fano lineshape) or semiconducting (Lorentzian lineshape).

The disorder-related mode (D peak) is a phonon mode at 1300 cm⁻¹ and involves the resonantly enhanced scattering of an electron via phonon emission by a defect that breaks the basic symmetry of the graphene plane. This mode corresponds to the conversion of a sp²-hybridized carbon to a sp³-hybridized on the surface. Intensity of D peak measures covalent bond made with the nanotube surface. This feature does not increase as a result of surfactant or hydronium ion adsorption on the nanotube surface.

Selective functionalization increases the intensity of the D peak due to formation of aryl-nanotube bond and decreases the tangential mode due loss of electronic resonance. These two effects are generally summed together as increase in their peaks ratio (D/G). RBM peaks of metallic nanotubes decay and the peaks corresponding to those of semiconducting nanotubes remain almost unchanged.[3][4]