Solid-state metathesis reactions have gained considerable attention over the past few years in the synthesis of refractory ceramics and semiconductors [33-36]. Unlike the long-lasting conventional solid-state reactions that encounter large diffusion barriers, rapid metathesis in solid-state can reach high temperature (>1000 0 C) on a very short time scale (<1 s) [37].

Very recently, solid-state synthesis of CNTs via metathesis has also been conducted using carbon halides (e.g., C2Cl6), lithium acetylide (Li2C2), and catalyst cobalt dichlo-ride (CoCl2). The chemical reaction can be given as [38]:

The theoretical maximum (exchange) reaction temperature of 2307 K was calculated, based on reaction enthalpy data of Eq. (10). The reactions are self-propagating and can be simply initiated with a heated filament. With 5 mol% (with respect to carbon) of cobalt dichloride added to the precursor compounds, both single- and multiwalled CNTs, along with graphite-encapsulated cobalt nanoparticles, have been synthesized. Some of the multiwalled nanotubes have a length up to 50 nm or more and possess a bamboolike structure (the morphology similar to that formed from vapor-phase syntheses!). Graphite-encapsulated cobalt nanoparticles, amorphous, and graphitic carbons are also found, noting that the latter two byproducts can easily be removed by washing in concentrated nitric acid. It has been found at lower catalyst concentrations (<5 mol%) that the yield of single-walled CNTs significantly reduces, while at higher Co concentrations (>5 mol%), the yield of graphite-encapsulated nanoparticles rises rapidly with no evidence for the formation of single-walled CNTs. Therefore, the 5 mol% CoCl2 is considered to be an optimal concentration for the formation of CNTs. Among the various catalysts tested (5 mol% FeCl3, 5 mol% NiCl2, 2-10 mol% CoCl2) for Eq. (1), it is noted that CoCl2 is the only case giving single-walled CNTs, although all of them can produce multiwalled CNTs and graphite-encapsulated metal nano-particles. Without a transition-metal catalyst, on the other hand, only graphitic and amorphous carbons can be formed.

The reaction temperature of these solid-state metathesis reactions can be further regulated. Calculation performed by the same group of researchers indicates that increasing the length of the carbon chain can lower the reaction temperatures. Furthermore, it has been shown that replacing chlorine with fluorine can also reduce the reaction temperature. In principle, hydrogen in substitution of chlorine can further decrease the reaction temperature. This theoretical prediction has been confirmed experimentally using a copolymer of poly(vinyl chloride) and poly(vinylidene chloride) with a 5 mol% of FeCl3 catalyst to form multiwalled CNTs [38]. It is believed that these reactions, once optimized, will represent an economic synthetic route using cheaper precursors under less-demanding reaction conditions for the CNTs production.

3.3. Thermal Reactions (Pyrolysis)

It has long been known that carbon can retain in carbon-containing ceramic solids after their thermal decompositions owing to insufficient oxidation reactions to CO and CO2 for the trapped carbon species under atmospheric conditions. For example, significant amounts of carbons were found in organic-compound intercalated hydrotalcites and sol-gel materials derived via organic routes after heat treatments before the discovery of CNTs [39-43]. As revealed later, these residual carbon phases may also be formed as fullerene-like carbons such as short multiwalled CNTs and polyhedral particles at presence of metal catalysts [44-46].

Carbon materials can be converted from polymeric materials via simple thermal treatments. In particular, these polymeric precursors can be fabricated into various shapes with the assistance of porous solid templates, as illustrated in Figure 1 for a hypothetical array of one-dimensional (1D) channels in synthesis of carbon nanotubes. As early as in 1989, inclusion of polyaniline filaments in zeolite molecular-sieves had been achieved [47]. In 1992, zeolite-Y and mordenite had been utilized as this kind of template materials for radical polymerization of acrylonitrile within their open pores and channels. The thus-prepared polyacry-lonitrile (PAN) can be converted into graphite-like conducting carbon filaments via pyrolysis at elevated temperatures

[48]. Using the similar solid-state approach, in-situ polymerization of acrylonitrile had also been carried out, producing PAN nanotubes within the pores of a nanoporous membrane

[49]. In a subsequent heat treatment, the thus-prepared PAN nanotubes underwent graphitization that gave aligned multiwalled CNTs, similar to those depicted in Figure 1.

Very recently, carbon 60 (C60) fullerene nanotubes have also been prepared from the template technique with 1D alumina channels with a density of 1011 channels/cm2 [50]. In this synthetic preparation, the alumina template was repeatedly dipped into a C60 containing toluene solution. After several dipping-and-drying cycles, the C60 deposited templates were heated at 500 0 C for 5 hours in an argon atmosphere and then cooled down to room temperature. The as-grown C60 nanotubes after removal of the template (with NaOH solutions) are well aligned (tube diameter 220-310 nm, and tube length 60 fim), showing a hexagonal long-range order in accordance with the original 1D channel array (e.g., Fig. 1). The surfaces of these tubes are smooth and clean, and the FTIR investigation shows four typical C60 absorption

Thermal Conversion of Polymeric Materials to CNTs

Figure 1. Schematic drawing of porous inorganic template growth of CNTs with prefilled organic materials in the one-dimensional channels and thermal conversion reactions.

Figure 1. Schematic drawing of porous inorganic template growth of CNTs with prefilled organic materials in the one-dimensional channels and thermal conversion reactions.

bands at 527, 577, 1182, and 1431 cm-1, respectively. The TEM and electron diffraction (ED) observations indicate that the C60 nanotubes are polycrystalline, and the structure of nanotubes is a mixed phase of face-centered cubic and hexagonal close-packed arrangements [50]. In view of its versatility, this approach should represent a promising route for formations of other nanotubular structures for various materials including polymer, inorganic, and organic compounds. In particular, this technique could be a general method for thermopolymerization and in-situ polymerization under controlled conditions.

Unlike the above cases, in which porous templates were used, nanotubular materials can also be prepared with negatype templates [51, 52]. In particular, micrometer-sized carbon tubes had been prepared by solid-state reactions with polymeric composite fibers that have a novel core-skin structure [53, 54], as described in Figure 2. The composite fibers were fabricated by a typical reaction coating method [55]. Under a nitrogen atmosphere (flow rate at 0.5 L/min), the composite fibers comprising a thermally removable polymer core [e.g., poly(ethylene terephthalate), PET] and a thermally more stable skin (e.g., conducting polymer: polypyr-role, PPy) were heated in a quartz tube oven from room temperature to 1000 °C (at a heating rate of 10 °C/min). Upon this thermal heating, the reactions were carried out at 1000 °C for 0 to 3 hours, respectively, prior to the thermally formed products being cooled down to ambient temperature. This pyrolysis method turned out to be very versatile. It had been used to prepare micrometer- or submicrometer-sized carbon tubes with controllable wall thickness in the range of smaller than 30 nm to a few micrometers and with a few centimeters in length by changing the thickness of the PPy skin coating as well as the diameter of the PET core fibers. More importantly, this method allows us to prepare well-organized two- or three-dimensional structures assembled by carbon tubes several centimeters in length via selecting or preparing appropriate woven polymeric templates [56]. It has been proven that the diameters of the PPy-skinned PET fiber composites are directly proportional to the number of coating treatments. For example, the corresponding weight gain of the composite samples was measured to be 0.66%, 9.3%, 18.7%, and 27.1% (with respect to the original uncoated PET fibers) when 1 to 4 cycles of PPy coating were applied, respectively. As shown in these weight data, the growth of PPy skin on the fresh PET surface (i.e., the first coating cycle) was significantly slower [54]. This weight-variation suggests that different formation mechanisms for different PPy coating layers might be involved.

PPy Shell on Carbon Tube PET Core

Figure 2. Synthetic process for making micrometer-sized carbon tubes using PET core templates.

PPy Shell on Carbon Tube PET Core

Figure 2. Synthetic process for making micrometer-sized carbon tubes using PET core templates.

By varying the concentrations of PPy and oxidant and reaction time, it is understood that most of PPy particles in the first coating cycle might have initiated and continued their growth in the solution phase, and only those close to the surfaces of PET fiber cores have chance to deposit via chemisorption and/or physisorption. Because of the competition between the solution phase and surface phase, on one hand, the resulting PPy in the first coating layer is thinner. Due to the lower oxidation potential of the first PPy layer, on the other hand, the subsequent coating or polymerization of pyrrole becomes easier. Scanning electron microscopic observations on these thermally treated samples reveal that the cores of composite fiber start to melt at a temperature between 230 and 290 °C and then decompose at 390 °C, leaving behind only carbon microtubes. Both the diameter and thickness of carbon tubes decrease continuously when the heating temperature is increased. Furthermore, since the PPy is thermally more stable than the PET, the resultant carbon tube walls are mainly derived from the PPy skin layers. With a recently developed method for fabrication of crystalline linear polyethylene (PE) nanofibers [57], carbon tubes with diameters in the nanometer regime can be further prepared using these nanometer-core fibers. However, it should be mentioned that the resultant carbon tubes prepared at 1000 °C are largely amorphous. With a further annealing at 1000-2400 °C, the disordered amorphous carbon tube walls can be gradually changed to a highly ordered graphitic phase with preferred orientation. Accordingly, in summary, the temperatures for these thermal reactions can be divided into three different zones: (i) 200-500 °C, formation of precursor tubes, (ii) 500-1700 °C, carbonization of the precursor tubes, and (iii) 1700-2400 °C, graphitization process [58] to increase the crystallinity of the carbon tubes.

Without the solid templates, synthesis of CNTs via direct thermal treatment of polymers (without using catalysts) can also be achieved [59]. The solid precursor used was a transparent brown resin prepared from anhydrous citric acid (HOOCCH2C(OH)(COOH)CH2COOH) and ethylene glycol (HOCH2CH2OH) with a molar ratio of 1:4. The resin was heat-treated at 400 °C for 8 hours in air on an alumina boat followed by natural cooling to room temperature. The CNTs obtained in such a way are multiwalled tubes with diameters ranging from 5 to 20 nm and lengths shorter than 1 ¡xm. In addition to the tubular morphology, small carbon graphitic particles and amorphous sheet-like carbons are also found. Regarding heating condition at 400 ° C, several kinetically competing chemical processes are brought to attention: formation reactions of CO, CO2, CNTs, graphitic particles, and amorphous sheet-like carbons. It is believed that the decomposition of polymer might generate dangling bonds of carbon for further reconstruction at this temperature. Therefore, the current CNT growth seems to take place as a nonequilibrium solid-state process. The simple synthetic conditions, such as the solid-state reaction, catalyst-free, and mild heat treatment in air atmosphere, are particularly attractive to mass-production of CNTs. It has also been confirmed that the CNTs can also be obtained under vacuum condition or with inert atmospheres for the similar reaction processes.

Solid-state synthesis of CNTs can also be simply conducted with organic compounds such as polymers or large hydrocarbons and transition metals as catalysts for direct thermal decomposition reactions. In this type of reactions, the CNTs are formed together with the original metal catalysts used in the synthesis. Without a further separation/purification, the resultant mixtures can be viewed naturally as the CNTs/metal nanocomposites. Nonetheless, the product CNTs can be easily separated from product mixtures via acid or thermal treatments after the solid-state catalytic reactions. It had been recently demonstrated that thermal decomposition of polyethylene with nickel catalyst is an efficient method to fabricate large quantities of multiwalled CNTs with fishbone-like morphologies [60]. The content of nickel nanoparticles in the final CNTs/metal nanocomposites is less than 15% by mass, and the mixing of both components is excellent. If desired, the nickel component can be completely removed by thermal processing of the nano-composites in vacuum at temperatures up to 2800 °C. The CNTs prepared are all multiwalled with lengths of several micrometers, inner channel diameter of 9-12 nm, and outer diameter of 40-50 nm. In particular, these CNTs consist of almost rectilinear sections with lengths of 100-300 nm turned with respect to one another, and their walls are composed in most cases of 40-65 tapered graphene layers with a taper angle along the tubes in the range of 16-35°. Interestingly, the dimensions and shapes of wider sections of the inside channel can be correlated to those of nickel nano-particles, which are located in most cases at the ends of the nanotubes. In order to have transport measurements, the mixture powders were cold pressed under high pressure of about 25 kBar and shaped into bars with dimensions of 1 x 2 x 3 mm3. It has been found that the electric resistivity (R) of these nanocomposites at room temperature is about 1 Hcm. The resistance was measured as a function of temperature down to the liquid-helium temperature in magnetic fields of up to 75 kOe. In the low-temperature range of 4.2-100 K, the resistivity of the composite changes according to the law of ln R a (To/T)1/3 where To ~ 7 K. The measured magnetoresistance can be interpreted in terms of two-dimensional variable range hopping conductivity. It has been assumed that the space between the inside and outside walls on the nanotubes acts as a two-dimensional medium. These CNTs with nested-cone morphologies are characterized by very high densities of electron states at the Fermi level of about 1021 eV-1cm-3, which is a typical value of metals [60].

Different from the above cases, in which organic solids were used as carbon sources, CNTs can also be synthesized from inorganic carbon-containing solids, such as SiC. For example, an aligned CNT film was formed and investigated from surface decomposition of S-SiC at 1700 ° C by in-situ transmission electron microscopy [61]. The degree of the orientation in the CNT film depends strongly on the S-SiC crystal planes; CNTs preferentially orient along [111] direction on the (111)s-SiC surface. Later, it was further found that the surface decomposition was progressed by an oxidation reaction with residual oxygen gas, based on the observations of high-resolution electron microscopy and electron energy-loss spectroscopy [62, 63]. Similar syntheses using single-crystal wafers of 6H-SiC with Si(0001) and C(0001) terminations were also carried out, aiming at a mechanistic understanding of the formation of aligned

CNTs [64]. As depicted in Figure 3, the SiC single wafers were heated at a temperature range of 1200-1700 °C for half an hour in a vacuum furnace with an electric resistance heater. On the C-face, vertically aligned CNTs 3-5 nm in diameter and 0.25 ¡m in length were formed after heating. In contrast, on the Si-face, graphite sheets (ca. 5 nm in thickness) parallel to the crystal plane were generated. The CNTs formed on the C-face were about 50 times thicker than the graphitic layers formed on the Si-face under the same decomposition conditions. Based on the experimental results, the difference in carbon products can be further explained. The graphitic sheets formed on the Si(0001) are attributable to an epitaxial growth as graphitic (0001) plane is parallel to the 6H-SiC(0001). On the basis of the two-and three-layer collapse mechanisms [65, 66], the area carbon density of the three condensed layers is around 3.67 x 1015 atoms cm-2, which is close to the carbon density of a graphite layer (3.80 x 1015 atoms cm-2). For the crystal plane with C-termination, carbon islands could be nucleated at impurity or defect sites, and develop into the larger particles of about 5 nm in size. There are two types of carbon islands, corresponding to spherical and flat caps for the final CNTs. The first one, spherical particles, would be graphi-tized from the surface. The second one, formed from two to five (flat) graphitic layers, would curve to connect with the carbon atoms on the surface. In particular, the diameter of the nanocaps depends on the SiC decomposition rate, the surface diffusion coefficient of carbon, the cohesive energy of carbon, and the flatness of the SiC surface [64]. As for the decomposing product SiO, the atomic oxidation mechanism of SiC is thought to be plausible. For example, gaseous SiO molecules could be continuously generated at the boundary between SiC and the graphite clusters. Nonetheless, the oxidation or decomposition process of SiC is thought to be rather complicated. The formation of Si—O—C bonding during the SiC oxidation has been investigated theoretically [67]. It has been realized that the atomic difference between the Si- and C-faces of SiC must have played an important role in the SiC decomposition.

In general, controlled synthesis of CNTs in solid state is much more difficult than that in vapor phase, especially

C-Face Termination

SiC Crystal lllllllllllllllllllllllllllllllllllllk^


Aligned CNTs

Figure 3. Growth of aligned CNTs using oriented SiC single-crystal wafers as solid precursors.

Aligned CNTs

Figure 3. Growth of aligned CNTs using oriented SiC single-crystal wafers as solid precursors.

for the preparation of single-walled CNTs. Very recently, this difficulty has been overcome, and aligned single-walled CNTs have been synthesized successfully in solid state. Using the Si-face of hexagonal silicon carbide (6H-SiC) at temperature above 1500 °C [68], two types of CNTs have been grown over the steps or terraces of the SiC. The first type, Y-shaped CNTs, results when the surface morphology is composed of terraces. The CNTs form a web-like network with a predominance of 120° for inter-bundle (or inter-tube) angle. The second type, T-shaped CNTs, is found when the surface is composed of parallel steps. In the latter case, the tubes align either along or perpendicular to the step edges which then give an inter-bundle (or inter-tube) angle of 90°. Most importantly, the CNTs prepared are single-walled with a very narrow distribution of diameters. With AFM manipulation, it is further demonstrated that the CNT bundles can be divided into several smaller tubes. In addition to this, it is also confirmed that the AFM manipulation can bring subsurface segments of tubes to the surface. After reanneal-ing the sample in ultra-high vacuum at 1300 °C, the CNTs can return to a well-ordered configuration. This observation indicates that the ordering of the CNTs is due to a diffusion process occurring after the CNT growth.

Regarding the formation mechanism of these flat-aligned CNTs, it is noted that when the Si termination of 6H-SiC is annealed at very high temperature, silicon will evaporate, leaving behind a carbon-rich surface. With atom-resolved STM technique, it is observed that this C-rich surface possesses the hexagonal C-lattice (i.e., graphite layers), and these initially formed graphite layers are not well-defined graphene sheets but more likely composed of small pieces of graphite layers with a high population of dangling bonds at their edges. Due to the thermodynamic driving force, the dangling carbon bonds can be saturated by folding of the graphene pieces. On the other hand, the thus-formed nanotube segments can act as seeds for the attachment of new carbon atoms, leading to the subsequent nanotube growth [68].

0 0

Post a comment