Elsevier

Carbon

Volume 49, Issue 12, October 2011, Pages 3803-3807
Carbon

Effect of bundling on the π plasmon energy in sub-nanometer single wall carbon nanotubes

https://doi.org/10.1016/j.carbon.2011.04.072Get rights and content

Abstract

Due to their unusual electronic and vibrational properties, single walled carbon nanotubes (SWCNTs) with sub-nanometer diameters d  0.5–0.9 nm have recently gained interest in the carbon community. Using UV–Vis–NIR spectroscopy and ultra-centrifugation, we have conducted a detailed study of the π plasmon energy (present at∼5–7 eV) in sub-nm SWCNTs as a function of the size of the bundle. We find that the energy of the π plasmon peak E varies with the bundle diameter Dh as E = (-0.023 eV)ln(Dh/do) + 5.37 eV, where do = 0.5 nm and corresponds to the smallest tube diameter.1 This is compared with the same data for HiPCo and Carbolex SWCNTs of larger diameter (1–1.4 nm) confirming a clear dependence of E on the bundle size, which is present in addition to the previously reported dependence of E on SWCNT diameter d.

Introduction

A single wall carbon nanotube (SWCNT) is a well known quasi one-dimensional system whose physical properties depend on the tube diameter and chirality [1]. The energy gap for the first (ES11) and second (ES22) pairs of van Hove singularities in the electron density of states for semiconducting SWCNTs range between 0.79–1.48 eV and 1.39–2.5 eV [2]. In metallic SWCNTs, EM11 transitions are present in the range 2.0–3.1 eV [3]. In addition to these inter band transitions Eii, UV–Vis–NIR and electron energy loss spectroscopy (EELS) have reported high energy features around 5.2 eV which correspond to collective excitations of the π electrons (π plasmon) [4], [5].2 In graphite, the carbon atoms exhibit sp2 hybridization in which each carbon atom is connected evenly to three carbons in the xy plane and a weak π bond is present in the z direction. This sp2 set forms the honeycomb lattice of a graphene sheet and the pz orbitals are responsible for the van der Waals bonding between graphene sheets. The free electrons in the pz orbitals are delocalized and behave like an electron cloud [6]. In a SWCNT, which is a rolled-up graphene sheet, the carbon–carbon bond becomes more sp3 like and the pz orbitals overlap due to the quasi 1D nature of nanotubes. This overlap becomes more significant as the diameter decreases below 1 nm [7]. The energy corresponding to the π plasmon energy depends on the nanotube diameter [8], [9] and the overlap integral [10]. Several reports find that a value of ∼2.9 eV for the overlap integral adequately explains the observed optical properties of SWCNTs with diameters in the range 1–3 nm [11], [12].

Recently, Rance et al. [8] reported the dependence of the π plasmon energy on the tube diameter in small SWCNT bundles derived from CoMoCAT, HiPCo, electric arc and chemical vapor deposition methods. However, their study did not address the effect of de-bundling on the π plasmon energy. Que [13] has proposed that different plasmon modes in a SWCNT bundle can couple due to inter tube interaction and result in a shift in the π plasmon energy. Here, we present UV–Vis–NIR data for sub-nm SWCNTs (d < 1 nm) with a view towards quantifying the effect of bundling on the π plasmon energy and providing an empirical relation between the π plasmon energy and the SWCNT bundle size.

Section snippets

Experimental section

Sub-nm SWCNT bundles were synthesized using Co:Mn catalyst in a molar ratio of 1:1 at three different temperatures (600, 700 and 800 oC). We refer to these samples as S1, S2 and S3. Their average tube diameters range from 0.5 to 0.9 nm [14]. Fig. SI1 shows transmission electron microscope images and the diameter distribution of as-synthesized sub-nm SWCNTs. In order to elucidate the dependence of π plasmon energy on bundle size, we adopted the method described by Arnold et al. [15] where in

Results and discussions

Fig.1 shows the UV–Vis–NIR spectra of as-prepared samples of S1, S2 and S3 before and after centrifugation for 15 and 60 min. Each of the absorption peaks corresponds to an electronic transition energy in semiconducting or metallic sub-nm SWCNTs., The peak positions are governed by the equations E11s=2ac-cγ0d, E22s=4ac-cγ0d and E11M=6ac-cγ0d, where ac-c is the carbon–carbon bond length, γ0 the overlap integral and d the tube diameter. After centrifugation for 15–60 min, the peaks corresponding to

Conclusion

In summary, the π plasmon energy dependence on the bundle size in sub-nm SWCNTs has been characterized using UV–Vis–NIR spectroscopy. DLS spectroscopy has been used to measure the hydrodynamic size of the SWCNT bundle and the dispersion quality is obtained from Zeta potential measurements. For the first time, an empirical relationship has been observed between the plasmon energy and nanotube bundle diameter. This relation is verified for Carbolex and HiPCo SWCNTs. The observed shift in the π

Acknowledgements

We would like to thank Department of Genetics and Biochemistry, Clemson University for use of their equipment and acknowledge Dr. Malcolm Skove at the Department of Physics, Clemson University for his constructive inputs. We gratefully acknowledge financial support from the US AFOSR Grant Number FA 9550-09-1-0384.

References (23)

  • A. Jorio et al.

    Structural (n,m) determination of isolated single-wall carbon nanotubes by resonant Raman scattering

    Phys Rev Lett

    (2001)
  • Cited by (0)

    1

    In order to make our empirical relation dimensionally correct, we divide the SWCNT bundle diameter Dh with the smallest tube diameter (do = 0.5 nm) present in the sub-nm SWNTs. This value for do is constant in the empirical relation and was determined from a careful Raman and photoluminescence measurements.

    View full text