Page 21 - Carbon Nanotubes

12 D. T. COLBERT and R. E. SMALLEY

inated as broken fragments of sintered nanotubes, the bias was controlled relative to an opposing electrode,
amount of remaining material reflects the degree of and if desired, reactive gases could be introduced.
sintering. Two classes of emission behavior were found. An
Our examinations of oxygen-purified deposits led inactivated state, in which the emission current in-
to construction of a model of nanotube growth in the creased upon laser heating at a fixed potential bias,
arc in which the nanotubes play an active role in sus- was consistent with well understood thermionic field
taining the arc plasma, rather than simply being a emission models. Figure la displays the emission cur-
passive product[2]. Imaging unpurified nanotube-rich rent as the laser beam is blocked and unblocked, re-
arc deposit from the top by scanning electron micros- vealing a 300-fold thermal enhancement upon heating.
copy (SEM) revealed a roughly hexagonal lattice of Etching the nanotube tip with oxygen while the tube
50-micron diameter circles spaced -50 microns apart. was laser heated to 1500°C and held at -75 V bias
After oxidative treatment the circular regions were seen produced an activated state with exactly the opposite
to have etched away, leaving a hole. More strikingly, behavior, shown in Fig. 2b; the emission current in-
when the deposit was etched after being cleaved ver- creased by nearly two orders of magnitude when the
tically to expose the inside of the deposit, SEM imag- laser beam was blocked! Once we eliminated the pos-
ing showed that columns the diameter of the circles sibility that species chemisorbed on the tip might be
had been etched all the way from the top to the bot- responsible for this behavior, the explanation had to
tom of the deposit, leaving only the intervening mate- invoke a structure built only of carbon whose sharp-
rial. Prior SEM images of the column material (zone 1) ness would concentrate the field, thus enhancing the
showed that the nanotubes there were highly aligned emission current. As a result of these studies[9], a dra-
in the direction of the electric field (also the direction matic and unexpected picture has emerged of the
of deposit growth), whereas nanotubes in the sur- nanotube as field emitter, in which the emitting source
rounding region (zone 2) lay in tangles, unaligned with is an atomic wire composed of a single chain of car-
the field[2]. Since zone 1 nanotubes tend to be in much bon atoms that has been unraveled from the tip by the
greater contact with one another, they are far more force of the applied electric field (see Fig. 2). These
susceptible to sintering than those in zone 2, resulting carbon wires can be pulled out from the end of the
in the observed preferential oxidative etch of zone 1. nanotube only once the ragged edges of the nanotube
These observations consummated in a growth layers have been exposed. Laser irradiation causes the
model that confers on the millions of aligned zone 1 chains to be clipped from the open tube ends, result-
nanotubes the role of field emitters, a role they play ing in low emission when the laser beam is unblocked,
so effectively that they are the dominant source of but fresh ones are pulled out once the laser is blocked.
electron injection into the plasma. In response, the This unraveling behavior is reversible and reproducible.
plasma structure, in which current flow becomes con-
centrated above zone 1, enhances and sustains the
growth of the field emission source-that is, zone 1 4. THE STRUCTURE OF AN OPEN NANOTUBE TIP
nanotubes. A convection cell is set up in order to al- A portion of our ongoing work focusing on sphe-
low the inert helium gas, which is swept down by col- roidal fderenes, particularly metallofullerenes, utilized
lisions with carbon ions toward zone 1, to return to the same method of production as was originally used
the plasma. The helium flow carries unreacted carbon in the discovery of fullerenes, the laser-vaporization
feedstock out of zone 1, where it can add to the grow- method, except for the modification of placing the
ing zone 2 nanotubes. In the model, it is the size and flow tube in an oven to create better annealing con-
spacing of these convection cells in the plasma that de- ditions for fullerene formation. Since we knew that at
termine the spacing of the zone l columns in a hex- the typical 1200°C oven temperature, carbon clusters
agonal lattice. readily condensed and annealed to spheroidal fuller-
enes (in yields close to 40%), we were astonished to
find, upon transmission electron micrographic exam-
3. FIELD EMISSION FROM AN ATOMIC WIRE
ination of the collected soots, multiwalled nanotubes
Realization of the critical importance played by with few or no defects up to 300 nm long[lO]! How,
emission in our arc growth model added impetus to we asked ourselves, was it possible for a nanotube
investigations already underway to characterize nano- precursor to remain open under conditions known to
tube field emission behavior in a more controlled man- favor its closing, especially considering the absence of
ner. We had begun working with individual nanotubes extrinsic agents such as a strong electric field, metal
in the hope of using them as seed crystals for con- particles, or impurities to hold the tip open for growth
trolled, continuous growth (this remains an active and elongation?
goal). This required developing techniques for harvest- The only conclusion we find tenable is that an in-
ing nanotubes from arc deposits, and attaching them trinsic factor of the nanotube was stabilizing it against
with good mechanical and electrical connection to closure, specifically, the bonding of carbon atoms to
macroscopic manipulators[2,8,9]. The resulting nano- edge atoms of adjacent layers, as illustrated in Fig. 2.
electrode was then placed in a vacuum chamber in Tight-binding calculations[l 1 J indicate that such sites
which the nanotube tip could be heated by applica- are energetically preferred over direct addition to the
tion of Ar+-laser light (514.5 nm) while the potential hexagonal lattice of a single layer by as much as 1.5 eV