blocking of the active surface. This is achieved by surface diffusion and the carbon
forms a skin around the main filament body, in step three. In the fourth step,
overcoating and deactivation of the catalyst and termination of tube growth takes
place. Oberlin et al. [101] proposed a mechanism in which bulk diffusion is insignificant and carbon is entirely transported around the particle by surface diffusion.
Dai et al. [49] proposed a mechanism wherein carbon forms a hemispherical graphene cap on the catalyst particle and the nanotubes grow from such a yarmulke.
The diameter of the nanotube is controlled by the size of the catalytic particle,
nanometer size particles yielding SWNTs. A crucial feature of this model is that it
avoids dangling bonds at all stages of growth. SWNTs produced by arc vaporization
may also be formed by the yarmulke mechanism.
A number of models have been proposed for the growth of MWNTs in the arc.
Endo and Kroto [102], based on the observation of C 2 ejection from C 60 in mass
spectrometry, suggest that tube formation processes are a consequence of the formation of fullerenes. Smalley [103], however, pointed out that only the growth of
outer layers of multi-walled tubes would be permitted by such a mechanism. Iijima
et al. [104] presented electron microscopy evidence for the open-ended growth of
carbon nanotubes and suggested that the termination of incomplete layers of carbon seen on the tube surface may arise because of the extension and thickening of
the nanotubes by the growth of graphite islands on the surfaces of existing tubes.
The nucleation of pentagons and heptagons on the open tube ends results in a
change in the direction of the growing tube and some novel morphologies, including one where the tube turns around 180
during the growth, have been observed. The growth is self-similar and fractal-like with the inner tubules telescoping out of the larger ones, with logarithmic scaling of the size.
Isotope scrambling experiments of Ebbesen et al. [105] show that under the
conditions of fullerene formation, the plasma has vaporized atoms of carbon.
Based on tube morphologies, a mechanism similar to that of Saito et al. [106],
wherein the carbonaceous material reaching the cathode anneals into polyhedral
particles, was suggested. Given the right conditions, the tip might open and continue to grow. Such a growth could occur from the outside inwards. Ebbesen et al.
[105] suggest the possibility of tubes forming directly from the closing of a large
graphene sheet. Such a suggestion gains credence from the simulations of Robertson et al. [107] who have examined the curling and closure of small graphitic
ribbons. Amelinckx et al. [108a] introduce the concept of a spatial velocity hodograph to describe the extrusion of a carbon tubule from a catalytic particle. The
model is consistent with the observed tubule shapes and explains how spontaneous plastic deformation of the tubule can occur. Amelinckx et al. [108b] propose a
model in which the graphene sheets can form both concentric cylinders and scrolltype structures. They also propose that the nanotubes nucleate from a large fullerene type dome. Maiti et al. [109] propose a model wherein nanometer-sized protrusions on the metal particle surface lead to the nucleation of SWNTs.
TEM examination of the carbonaceous products obtained from the pyrolysis of
hydrocarbons and organometallic precursors indicates that the size of the catalyst
particle plays an important role (see Figure 8.10).
8.2 Carbon Nanotubes 223
forms a skin around the main filament body, in step three. In the fourth step,
overcoating and deactivation of the catalyst and termination of tube growth takes
place. Oberlin et al. [101] proposed a mechanism in which bulk diffusion is insignificant and carbon is entirely transported around the particle by surface diffusion.
Dai et al. [49] proposed a mechanism wherein carbon forms a hemispherical graphene cap on the catalyst particle and the nanotubes grow from such a yarmulke.
The diameter of the nanotube is controlled by the size of the catalytic particle,
nanometer size particles yielding SWNTs. A crucial feature of this model is that it
avoids dangling bonds at all stages of growth. SWNTs produced by arc vaporization
may also be formed by the yarmulke mechanism.
A number of models have been proposed for the growth of MWNTs in the arc.
Endo and Kroto [102], based on the observation of C 2 ejection from C 60 in mass
spectrometry, suggest that tube formation processes are a consequence of the formation of fullerenes. Smalley [103], however, pointed out that only the growth of
outer layers of multi-walled tubes would be permitted by such a mechanism. Iijima
et al. [104] presented electron microscopy evidence for the open-ended growth of
carbon nanotubes and suggested that the termination of incomplete layers of carbon seen on the tube surface may arise because of the extension and thickening of
the nanotubes by the growth of graphite islands on the surfaces of existing tubes.
The nucleation of pentagons and heptagons on the open tube ends results in a
change in the direction of the growing tube and some novel morphologies, including one where the tube turns around 180
during the growth, have been observed. The growth is self-similar and fractal-like with the inner tubules telescoping out of the larger ones, with logarithmic scaling of the size.
Isotope scrambling experiments of Ebbesen et al. [105] show that under the
conditions of fullerene formation, the plasma has vaporized atoms of carbon.
Based on tube morphologies, a mechanism similar to that of Saito et al. [106],
wherein the carbonaceous material reaching the cathode anneals into polyhedral
particles, was suggested. Given the right conditions, the tip might open and continue to grow. Such a growth could occur from the outside inwards. Ebbesen et al.
[105] suggest the possibility of tubes forming directly from the closing of a large
graphene sheet. Such a suggestion gains credence from the simulations of Robertson et al. [107] who have examined the curling and closure of small graphitic
ribbons. Amelinckx et al. [108a] introduce the concept of a spatial velocity hodograph to describe the extrusion of a carbon tubule from a catalytic particle. The
model is consistent with the observed tubule shapes and explains how spontaneous plastic deformation of the tubule can occur. Amelinckx et al. [108b] propose a
model in which the graphene sheets can form both concentric cylinders and scrolltype structures. They also propose that the nanotubes nucleate from a large fullerene type dome. Maiti et al. [109] propose a model wherein nanometer-sized protrusions on the metal particle surface lead to the nucleation of SWNTs.
TEM examination of the carbonaceous products obtained from the pyrolysis of
hydrocarbons and organometallic precursors indicates that the size of the catalyst
particle plays an important role (see Figure 8.10).
8.2 Carbon Nanotubes 223