Electrochemical Doping of Chirality-Resolved Carbon Nanotubes Ladislav Kavan,* ,† Martin Kalba ´ c ˇ , ‡ Marke ´ ta Zukalova ´ , † and Lothar Dunsch ‡ J. HeyroVsky ´ Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, Dolejs ˇkoVa 3, CZ-182 23 Prague 8, Czech Republic, and Leibniz Institute of Solid State and Materials Research, Helmholtzstr. 20, D-01069 Dresden, Germany ReceiVed: June 1, 2005; In Final Form: August 6, 2005 Raman spectra of electrochemically charged single-wall carbon nanotubes (HiPco) were studied by five different laser photon energies between 1.56 and 1.92 eV. The bands of radial breathing modes (RBM) were assigned to defined chiralities by using the experimental Kataura plot. The particular (n,m) tubes exhibit different sensitivity to electrochemical doping, monitored as the attenuation of the RBM intensities. Tubes which are in good resonance with the exciting laser exhibit strong doping-induced drop of the RBM intensity. On the other hand, tubes whose optical transition energy is larger than the energy of an exciting photon show only small changes of their RBM intensities upon doping. This rule presents a tool for analysis of mixtures of single-walled carbon tubes of unknown chiralities. It also asks for a re-interpretation of some earlier results which were reported on the diameter-selectivity of doping. The radial breathing mode in strongly n- or p-doped nanotubes exhibited a blue-shift. A suggested interpretation follows from the charging-induced structural changes of SWCNTs bundles, which also includes a partial de-bundling of tube ropes. 1. Introduction Raman spectra of single-wall carbon nanotubes (SWCNTs) are strongly resonant-enhanced via the optical transition between van Hove singularities (vHs). 1 This also provides a method for mapping of electronic structure and optical properties of SWCNTs. Electrochemical charging of SWCNTs allows further upgrade of this methodology by defined tuning of the electronic structure of SWCNTs, thus addressing the doping- induced changes of the Raman spectra in-situ. 2-8 The most obvious effect upon electrochemical doping is the attenuation of Raman intensities when the optical transitions between vHs are bleached, either by inserting extra electrons into the conduction band singularities (n-doping) or holes into the valence band singularities (p-doping). 2-8 Samples of SWCNTs, produced in a variety of preparative methods, contain both metallic and semiconducting tubes of variable chiral indices (n,m), and these tubes are usually assembled in bundles. Except for a few studies on one isolated SWCNT, 2,3 all spectroelectrochemical works carried out to date had supplied convoluted information about a mixture of tubes. 4-8 However, recent advances in Raman, optical, and photolumi- nescence spectroscopy of surfactant-wrapped SWCNTs revealed accurate relations between the optical transition energies and the Raman radial breathing mode (RBM). 9-12 The correlation of RBM frequencies with the empirical values of optical transition energies (the so-called “experimental Kataura plot”) 9-11 allows a unique (n,m) assignment, particularly for narrow semiconducting tubes of diameters between about 0.7 and 1.2 nm. The latter occur, for example, in the tubes made by a high- pressure catalytic decomposition of carbon monoxide (HiPco). Charge-transfer processes, employing inorganic 13 or organic 14 acceptor molecules as well as O 2 -mediated acid-basic reac- tions, 15 have been studied in detail with aqueous solutions of surfactant-wrapped HiPco tubes. Optical absorbance, Raman, and fluorescence spectra of (n,m)-resolved semiconducting tubes provided consistent evidence that the charge-transfer is chirality- selective. In particular, semiconducting tubes show an increase of the redox potential with increasing band gap. 13-15 Within the simplest tight-binding approximation, the energy separation of ith singularities, ΔE ii , is inversely proportional to the tube diameter, d: where γ 0 is the nearest-neighbor overlap integral (≈2.5 eV) and a CC is the CC bond length (≈142 pm). Due to curvature and many-body effects, the tight-binding model is perturbed, especially for narrow tubes (d ≈ <1.2 nm). 12 Nevertheless, for the discussion presented below, we can neglect the deviations of ΔE ii from the predicted 1/d behavior (eq 1) for smaller d. Spectroelectrochemistry on one isolated surfactant-wrapped tube 2,3 and chemical oxidation of tubes in solution 13,14 indicate that the position of the Fermi level (and band edges) scale linearly with ΔE ii . 2,3,14 In other words, narrow tubes should be stronger reductants (easily oxidizable), whereas wide tubes should be stronger oxidants (easily reducible) (eq 1). There are less detailed data on redox reactions of surfactant- free tubes in the solid phase, as these samples can hardly be investigated by fluorescence spectroscopy. Hence, the method of first choice is Raman spectroscopy, which, however, provides information on the band energetics only indirectly, via the resonance enhancement. Doping is carried out by the reaction of solid HiPco tubes with gaseous 16-18 or liquid 19 redox-active molecules, as well as electrochemically. 7,8 After charge-transfer, the compensating counterion is assumed to be located at the outer surface of the tube, but small gaseous molecules can also penetrate into the interior of a tube. 20 Recently, Iijima et al. 21 * Corresponding author. E-mail: Kavan@jh-inst.cas.cz. † J. Heyrovsky ´ Institute of Physical Chemistry, Academy of Sciences of the Czech Republic. ‡ Leibniz Institute of Solid State and Materials Research. ΔE ii ) 2iγ 0 a CC d (1) 19613 J. Phys. Chem. B 2005, 109, 19613-19619 10.1021/jp052910e CCC: $30.25 © 2005 American Chemical Society Published on Web 10/05/2005