Electronic Properties of Carbon Nanotubes
192
of graphene only, the energy gap between q
th
van Hove singularity in valence and
conduction bands for metallic tubes[126] is
3
1,2.3,…. ) and
for semiconducting tubes as discussed in Section 2. In the case of metallic tubes,
the linear crossing of π and π* bands near the Fermi level yields a small constant DOS
leading to sharp van Hove singularities at energies away from the Fermi energy. In contrast,
for semiconducting tubes, the DOS at Fermi energy is zero with sharp van Hove
singularities away from the Fermi energy [92,127].
The electronic structures of the SWBNNTs were theoretically investigated by Rubio et al.
[56] using the tight binding approximation (TBA). All SWBNNTs were found to be
semiconducting materials with band gaps larger than 2 eV. BNNTs with larger diameters
have a larger band gap, with a saturation value corresponding to the band gap of a
hexagonal boron nitride (BN) as calculated by Rubio et al.[56] or see the Figure 4 in Ref.[57].
The subsequent studies based on local-density-functional (LDA) calculations proved that it
is energetically more favorable to fold a hexagonal boron nitride sheet into a BNNT than to
forma CNT from a graphite sheet [58]. This observation is also supported by our
calculations at B3LYP/6-31G level of DFT as discussed in Section 3. Based on the band-
folding analysis, SWBNNTs are semiconductors with large-gap for the zigzag- and
armchair-SWBNNTs. This is due to the strong hybridization effects may take place because
tube curving reduces the BNNT’s band gap remarkably. For the (n, 0) BNNTs, with n > 12
the hybrid state was found not to play any role in determining tube gaps. The gaps are
steady at around 4 eV as indicated by LDA. In another calculation using density functional
theory, the band gaps of BNNTs were eventually saturated at 5.03 eV with an increase in
diameter [128,129,130,131,132]. Based our calculated dipole allowed vertical electronic
transitions, for the (n, 0)-BNNTs with n > 11, the gabs are steady around 5.9 eV, see Figure
3.6.1B and Table 3.6.4. For the (n,n)-SWBNNTs, with n>10, the band gaps slightly increases
with increasing tube diameter and seem to be steady at around 5.84 eV according to our
DFT calculations as seen Figure 3.6.1B and Table 3.6.4. It should be noted that there are
many forbidden electronic transition lie below the allowed one. Based on our experience on
the calculated electronic transitions for many organic compounds at B3LYP level and at the
BLYP level of DFT for the transition metals [95], the calculated value of the electronic
transitions are somewhat higher than their experimental values, as much as 0.4 eV.
Therefore, the calculated values of the electronic band gabs for the nanotubes might be
reasonable.
Various methods have been developed to tune the electronic structure of BNNTs is nicely
discussed by Chunyi Zhi et al.[57], for instance, applying electric field
[133,134,135,136,137,138] or strain [139], or chemical methods, like doping [140,141,142,67],
introducing defects [143,144,145] or surface modification [60,146,147]. The methods and
species chosen are summarized in Table 2 in Ref. 57. Physical methods could directly reduce
the band gap of BNNTs, while chemical methods were found to tune the band gap by
introducing localized energy levels inside the gap [57]. For example, by applying an 2 V/nm
electric field, the band gap of a (12, 12) BNNT was directly reduced from 4.5 eV to about 2.5
eV [135], while F-doping induced an unoccupied localized state in the gap of BNNTs [148].
In addition, chemical modification methods of BNNTs were sometimes found to enrich the
properties of BNNTs, for instance, polarization field was induced by chemical adsorption
[149] or ferromagnetism appeared in doped BNNTs.