Catalytic Growth Mechanisms for Single Shell Nanotubes

All these classical and quantum simulations, described above, may explain why single-shell nanotubes do not grow in the absence of transition metal catalysts. However, the role played by these metal atoms in determining the growth has been inaccessible to direct observation and remains controversial. The catalytic growth mechanism has not yet been clearly understood, but plausible suggestions include metal atoms initially decorating the dangling bonds of an open fullerene cluster, thus preventing it from closing. As more carbon atoms collide with this metal-decorated open-carbon cluster, they are inserted between the metal and the existing carbon atoms in the shell [10].

Planar polyyne rings have also been proposed to serve as nuclei for the formation of single-wall tubes, whose diameter would be related to the ring

Fig. 9. Spontaneous closure of two single-walled nanotubes: (a) (10,0) zigzag tube and (b) (5,5) armchair tube. The notation of [29] is adopted. Both nanotubes have a fully reconstructed closed-end configuration with no residual dangling bonds [22,23]. The present systems contains 120 carbon atoms (large white spheres) and 10 hydrogen atoms (small dark grey spheres) used to passivate the dangling bonds on one side of the two clusters (bottom). The other low coordinated carbon atoms (dangling bonds) are represented as light grey spheres on the top of the structure

Fig. 9. Spontaneous closure of two single-walled nanotubes: (a) (10,0) zigzag tube and (b) (5,5) armchair tube. The notation of [29] is adopted. Both nanotubes have a fully reconstructed closed-end configuration with no residual dangling bonds [22,23]. The present systems contains 120 carbon atoms (large white spheres) and 10 hydrogen atoms (small dark grey spheres) used to passivate the dangling bonds on one side of the two clusters (bottom). The other low coordinated carbon atoms (dangling bonds) are represented as light grey spheres on the top of the structure size [31]. In this model, the starting materials for producing nanotubes are monocyclic carbon rings, acting as nanotube precursors and ComC„ species acting as catalysts. The composition and structure of the Co carbide cluster are undetermined, but the cluster should be able to bond to Cn and/or to add the Cn to the growing tube. The helical angle of the single-shell tube produced is determined by the ratio of cis to trans conformations of the first benzene ring belt during the growth initiation process. The present model may explain the formation of larger tubes upon the addition of S, Bi, or Pb to the Co catalyst [32], as these promoters modify the growth at the nucleation stage and may stabilize larger monocyclic rings, providing the nuclei necessary to build larger diameter tubes [7].

Since transition metals have the necessary high propensity for decorating the surface of fullerenes, a layer of Ni/Co atoms, adsorbed on the surface of C60, also provides a possible catalyst template for the formation of single-walled nanotubes of a quite uniform diameter [33]. In this scenario, the metal coated fullerene acts as the original growth template inside the cylinder (Fig. 10). Once the growth has been initiated, nanotube propagation may occur without the particle. The optimum diameter of the nanotube product can thus be predicted to be the sum of the diameter of the fullerenes (0.7 nm) and two times the metal ring distances (2 x 0.3-0.35 nm), which is close to the observed value of 1.4 nm [5].

Another possibility is that one or a few metal atoms sit at the open end of a precursor fullerene cluster [5], which will be determining the uniform diameter of the tubes (the optimum diameter being determined by the competition between the strain energy due to curvature of the graphene sheet and the dangling bond energy of the open edge). The role of the metal catalyst is to prevent carbon pentagons from forming by "scooting" around the growing edge (Fig. 11).

Static ab initio calculations have investigated this scooter model and have shown that a Co or Ni atom is strongly bound but still very mobile at the growing edge [34]. However, the metal atom locally inhibits the formation of pentagons that would initiate dome closure. In addition, in a concerted ex-

Fig. 10. Schematic ball-and-stick representation of a transition metal surface decorated fullerene (Ceo) inside a open-ended (10,10) carbon (white spheres) nanotube. The Ni and Co atoms (large dark spheres) adsorbed on the C60 surface are possible agents for the creation of single-walled nanotubes of uniform diameter [33]

Fig. 11. View of a (10,10) armchair nanotube (white ball-and-stick atomic structure) with a Ni (or Co) atom (large black sphere) chemisorbed onto the open edge [5]. The metal catalyst keeps the tube open by "scooting" around the open edge, insuring that any pentagons or other high energy local structures are rearranged to hexagons. The tube shown has 310 C atoms change mechanism, the metal catalyst assists incoming carbon atoms in the formation of carbon hexagons, thus increasing the tube length. With a non-vanishing concentration of metal atoms in the atmosphere, several catalyst atoms will eventually aggregate at the tube edge, where they will coalesce. The adsorption energy per metal atom is found to decrease with increasing size of the adsorbed cluster [34]. The ability of metal clusters to anneal defects is thus expected to decrease with their increasing size, since they will gradually become unreactive and less mobile. Eventually, when the size of the metal cluster reaches some critical size (related to the diameter of the nanotube), the adsorption energy of the cluster will decrease to such a level that it will peel off from the edge. In the absence of the catalyst at the tube edge, defects can no longer be annealed efficiently, thus initiating tube closure. This mechanism is consistent with the experimental observation (see Fig. 12) that no "observable" metal particles are left on the grown tubes [5]. This also suggests that too high a concentration of the metal catalyst will be detrimental to the formation of long nanotubes.

Although the scooter model was initially investigated using static ab initio calculations [34], first-principles molecular dynamics simulations were also performed [35] in order to study the growth of single-wall tubes within a scheme where crucial dynamical effects are explored by allowing the system to evolve free of constraints at the experimental temperature. Within such a simulation at 1500 K, the metal catalyst atom is found to help the open end of the single-shell tube close into a graphitic network which incorporates the catalyst atom (see Fig. 13). However, the cobalt-carbon chemical bonds are frequently breaking and reforming at experimental temperatures, providing the necessary pathway for carbon incorporation, leading to a closed-end catalytic growth mechanism.

1 nm

Fig. 12. Transmission electron micrographs of single-shell nanotube ends. No particles of metals are observed at the nanotube caps [17]

Fig. 13. Catalytic growth of a (6,6) armchair single-walled nanotube. The notation of [29] is adopted. The metal catalyst atom cannot prevent the formation of pentagons, leading to tube closure at experimental temperatures (1500 K) [35]. The present systems contains 1 cobalt atom (large black sphere), 120 carbon atoms (white spheres) and 12 hydrogen atoms (small dark grey spheres) used to passivate the dangling bonds on one side of the two clusters (bottom). The other low coordinated carbon atoms (dangling bonds) are represented as light grey spheres on the top of the structure (left)

Fig. 13. Catalytic growth of a (6,6) armchair single-walled nanotube. The notation of [29] is adopted. The metal catalyst atom cannot prevent the formation of pentagons, leading to tube closure at experimental temperatures (1500 K) [35]. The present systems contains 1 cobalt atom (large black sphere), 120 carbon atoms (white spheres) and 12 hydrogen atoms (small dark grey spheres) used to passivate the dangling bonds on one side of the two clusters (bottom). The other low coordinated carbon atoms (dangling bonds) are represented as light grey spheres on the top of the structure (left)

This model, where the Co or Ni catalyst keeps a high degree of chemical activity on the nanotube growth edge, clearly differs from the uncatalyzed growth mechanism of a single-wall nanotube discussed above, which instan taneously closes into an chemically-inert carbon dome. The model, depicted in Fig. 13, supports the growth by chemisorption from the vapor phase, as proposed long ago for carbon filaments by Baker et al. [36], Oberlin et al. [8], and Tibbetts [9], which adopts the concepts of the vapor-liquid-solid (VLS) model introduced in the 1960s to explain the growth of silicon whiskers [37]. In the VLS model, growth occurs by precipitation from a super-saturated catalytic liquid droplet located at the top of the filament, into which carbon atoms are preferentially absorbed from the vapor phase. From the resulting super-saturated liquid, the solute continuously precipitates, generally in the form of faceted cylinders (VLS-silicon whiskers [37]) or tubular structures [9]. Tibbetts also demonstrates that it is energetically favorable for the fiber to precipitate with graphite basal planes parallel to the exterior planes, arranged around a hollow core, thus forming a wide diameter 1|j.m) multiwall tube.

Although the VLS model for catalyst-grown carbon fibers is a macroscopic model based on the fluid nature of the metal particle which helps to dissolve carbon from the vapor phase and to precipitate carbon on the fiber walls, the catalytic growth model for the single-wall nanotubes can be seen as its analogue, with the only difference being that the catalytic particle is reduced to one or a few metal atoms in the case of the single-wall nanotube growth. In terms of this analogy, the quantum aspects of a few metal atoms at the growing edge of the single-shell tube has to be taken into account. In the catalytic growth of single-shell nanotubes, it is no longer the "fluid nature" of the metal cluster (VLS model), but the chemical interactions between the Co or Ni 3d electrons and the n carbon electrons, that makes possible a rapid incorporation of carbon atoms from the plasma. The cobalt 3d states increase the DOS near the Fermi level, thus enhancing the chemical reactivity of the Co-rich nanotube tip [35].

The catalytic particle is not frequently observed at the tube end [5], although some experiments [38] report the presence of small metal particles on the walls of nanotube bundles (Fig. 14).

Precisely because of the enhancement of the chemical activity due to the presence of metal catalyst particles at the experimental growth temperature, the Co-C bonds are frequently reopened and the excess of cobalt could be ejected from the nanotube tip. Moreover, single-layer tubes have also been produced by pre-formed catalytic particles [39], which are attached to the tube end and thus are closely correlated with the tube diameter size (from 1 to 5nm), demonstrating the validity of the VLS model on a nanometer scale. When the metal particle is observable [39], the metal-catalytic growth of carbon nanotubes is thus believed to proceed (in analogy to the catalyst-grown carbon fibers) via the solvation of carbon vapor into tiny metal clusters, followed by the precipitation of excess carbon in the form of nanotubes.

The presence of any remaining metal catalyst atoms at the nanotube tip (even in very small - undetectable amounts) cannot be excluded. The presence of such ultra-small catalyst particles, which is certainly not easy to es-

Fig. 14. A high-resolution electron micrograph showing raft-like bundles of singlewalled carbon nanotubes. Most of the nanotubes form arch-like structures and they terminate at or originate from carbon blocks composed of mostly disordered cage-like fullerene molecules [38]. Some of the nanotube bundles are terminated in metal particles of a few nanometers in diameter. Smaller metal particles were also observed on the walls of the nanotube bundles [38]

Fig. 14. A high-resolution electron micrograph showing raft-like bundles of singlewalled carbon nanotubes. Most of the nanotubes form arch-like structures and they terminate at or originate from carbon blocks composed of mostly disordered cage-like fullerene molecules [38]. Some of the nanotube bundles are terminated in metal particles of a few nanometers in diameter. Smaller metal particles were also observed on the walls of the nanotube bundles [38]

tablish experimentally, should strongly influence the field emission properties of these single-shell tubes [40] and could explain some field emission patterns observed at the nanotube tip [41]. Magnetic susceptibility measurements and magnetic STM could be used to investigate the presence of metal catalysts at the nanotube tip, as such experimental techniques are sensitive to tiny amounts of magnetic transition metals. At this stage, a better experimental characterization method for the atomic structure of single-walled nanotube tips is required.

0 0

Post a comment