J I I 1 J

Fig. 20.16. Sequence of steps (deposition, exposure, development, and pattern transfer) used in fullerene photolithography in which the CWJ acts as a negative photoresist [20.68].

tation and was prepared as a hydrophilie surface (see §9.2.1). Some pressure was needed to bond the pair of wafers together. Interface energies of ~40 erg/cm2 and ~20 erg/cm2 were reported for the C60-SiO2 and C60-Si interfaces and were measured by the crack length generated by inserting a razor blade at the bonding interface [20.42]. Examination of the debonded surfaces did not show adhering C60 layers, indicating a lower cohesive energy at the interface than between C60 molecules. It was also found that two C60-covered Si wafers can easily be bonded to each other [20.42], It was further speculated that a thin fullerene film could be used to bond selected smooth surfaces to one another or to bond an Si wafer to certain other smooth surfaces [20.70]. Since the strongest bonding is at the C60-Si interface, a single C60 monolayer would be expected to show bonding action to Si.

20.2.7. Passivation of Reactive Surfaces

Using Auger electron spectroscopy (see §17.3) and temperature-programmed desorption (see §17.5), it has been demonstrated that a monolayer of C60 is sufficient to passivate a clean aluminum surface from room temperature up to 600 K against oxidation by water vapor, as shown in Fig. 20.17. In this figure, clear Auger lines are seen for aluminum (68 eV) and for carbon (272 eV), but essentially none are seen for oxygen (503 eV) after exposure of Al-C60 to about 340 monolayers of water vapor [20.71]. The amount of desorption of C60 multilayers to eventually yield monolayer coverage is determined by the intensity of the carbon to aluminum Auger lines in Fig. 20.17. Above ~700 K, the C60 molecules were found to become dissolved in the near-surface region of the aluminum substrate. Even in the dissolved phase, some surface passivation of the A1 surface by the presence of dissolved C60 was reported [20.71], In addition to inhibiting oxidation of the highly reactive A1 surface, a monolayer of C60 has been shown to protect the highly reactive Si (111)7 x 7 and Si (100)2 x 1 surfaces against oxidation [20.71-73],

20.2.8. Fullerenes Used for Uniform Electric Potential Surfaces

For some electronics applications, surfaces with a uniform electrical potential are required. In the past, graphite and gold surfaces have been used to provide equipotential surfaces. It has been found that by applying a thin (~10 A) deposit of germanium to a substrate (such as amorphous carbon, -y-irradiated NaCl or polished phosphorus bronze) and subsequently depositing an overlayer of ~100 A of fullerenes, a good equipotential surface is obtained [20.62]. The dangling bonds of the thin Ge interface layer would be expected to facilitate bonding to both the substrate and the fullerenes (see §17.9.3). Amorphous fullerene growth and uniform surface coverage of C60 on the Ge interface layer would be expected at room temperature, because of the low mobility and high sticking coefficient of the fullerenes to the Ge dangling bonds. The demonstration of this fullerene application was made with 85% C60 and 15% C70. Fullerenes are useful for providing equipotential surfaces on metals because fullerenes are chemically stable and easy to deposit in vacuum [20.62],

Electron energy (eV)

Fig. 20.17. A series of Auger electron spectra showing the passivation of aluminum by multilayer and single layer coverage of CM. The traces are displayed in chronological order from the bottom to the top. The bottom trace is for the A1 surface after it was sputter cleaned and annealed at 800 K. The next trace up is for the clean surface exposed to Qq for a sufficient time to form a multilayer coverage. The Langmuir (L) denotes a unit of exposure with 1L = 10~6 torr sec, so that with a sticking coefficient of unity, 1L corresponds to 1 monolayer coverage. Each succeeding trace is for the sample receiving the indicated treatment and all previous treatments [20.71],

Electron energy (eV)

Fig. 20.17. A series of Auger electron spectra showing the passivation of aluminum by multilayer and single layer coverage of CM. The traces are displayed in chronological order from the bottom to the top. The bottom trace is for the A1 surface after it was sputter cleaned and annealed at 800 K. The next trace up is for the clean surface exposed to Qq for a sufficient time to form a multilayer coverage. The Langmuir (L) denotes a unit of exposure with 1L = 10~6 torr sec, so that with a sticking coefficient of unity, 1L corresponds to 1 monolayer coverage. Each succeeding trace is for the sample receiving the indicated treatment and all previous treatments [20.71],

20.3. Materials Applications

In this section we present several current examples of situations where fullerenes offer promise for the synthesis of new materials and for better preparation methods for well-known materials.

20.3.1. Enhanced Diamond Synthesis

The possible use of C60 for the fabrication of industrial diamonds offers another area for possible fullerene applications. It is found that when non-hydrostatic pressures (in the range of 20 GPa) are applied rapidly at room temperature to Qq, the material is quickly transformed into bulk polycrys-talline diamond at high efficiency [20.74]. It is believed that the presence of pentagons in the C60 structure promotes the formation of sp3 bonds during the application of high anisotropic stress.

Normally, for diamond films to grow from a mixture of gaseous CH4 and H2, the substrate surface (usually Si) must be pretreated with diamond grit polish or must contain small diamond seeds. Enhanced nucleation of a high density of diamond crystallites on an Si substrate has been reported through the deposition of a 500-1000 A C70 film, activation of the film by positive ion bombardment, followed by chemical vapor deposition (CVD) growth in a microwave plasma reactor. A base layer of ion-activated C70 was found to be more effective in promoting sp3 bonding [20.75,76] than a similarly treated C60 buffer layer.

One possible explanation for the enhancement of the diamond nucleation process by the presence of fullerenes comes from the STM studies of C60 on Si (111) surfaces (see §17.9.3) [20.77], which show the C60 cages to burst above 1020 K. The microwave discharge described above may serve to clean the Si surface, thereby allowing strong bonding of the C60 to the Si surface. With this strong Si-C bond, the surface containing C60 can be heated to a high enough temperature to evaporate all the C60 except for the surface monolayer, which opens up above 1000 K, leaving active carbon sites for diamond nucleation [20.78].

20.3.2. Enhanced SiC Film Growth and Patterning

Several groups have studied the interaction of C60 with Si surfaces at high temperatures [20.77-85], and several groups have investigated the deposition of C60 on hot Si substrates as a method for enhancing the growth of SiC films [20.77-81,84]. Film growth occurs by first nucleating the SiC growth at ~1250 K and then by Si diffusion to the SiC interface to react with the Qq molecules that arrive at the free surface in a continuous flux. The Si diffusion coefficient through the SiC film is estimated to be between

10-13 and 10"12 cm2/s at a substrate temperature between 1100 and 1170 K [20.71]. SiC growth by this method on Si (previously cleaned at 1300 K) has been demonstrated on both Si (111) and Si (100) substrates heated to temperatures in the 950-1250 K range. Epitaxial SiC (100) lxl films up to 1 /urn in thickness can be grown in the cubic phase, as shown by x-ray diffraction and low-energy electron diffraction (LEED) patterns, while Auger spectroscopy was used to verify the SiC stoichiometry [20.86].

By exploiting the strong C60-Si interaction on a clean Si surface and the weak C60-Si interaction on a Si02 surface in comparison to the Qq-Qq bonding itself (see Table 17.2), it is possible to pattern the SiC film which grows readily on the clean Si surface but does not grow on the Si02 surface. To prepare a patterned SiC film, the SiC growth is carried out on an Si surface that was previously patterned by standard lithographic techniques with contrast provided by the regions of clean Si and clean Si02, and patterned SiC film growth to ~1 /xm thickness is possible. If Si rather than Si02 is desired in the regions where SiC growth has not occurred, the Si02 material can be removed with an HF etch. If a free-standing patterned SiC film is desired, liftoff of the SiC film from the substrate can be accomplished using an atomic force microscope (AFM) tip with a force of ~10~8 N between the tip and the surface, utilizing the weak adhesion and the large interface strain of thick (~1 /xm) SiC films on Si due to the ~20% lattice mismatch between SiC and Si. The AFM tip can also be used to manipulate and position the SiC films on a substrate. The SiC films produced through C60 precursors have a density of ~83% of the ideal density for SiC. Patterned SiC films may find use in high-temperature electronics and in micromechanical systems (MEMS) applications utilizing their high strength modulus (310 GPa), hardness (26 GPa), high thermal conductivity, and low coefficient of friction (one half to one third that of Si) [20.87].

20.3.3. Catalytic Properties of C60

A few reports have appeared in the literature referring to the catalytic activity of C60 [20.88-91]. Most of these reports have shown that C60 and C70 exhibit, at best, moderate catalytic activity. The choice of model compound reactions to evaluate their catalytic reactivity has been motivated by the tendency of fullerenes to form anions and hydrogenated adducts. Thus C60 has been tested as a catalyst for intermolecular hydrocarbon coupling (i.e., the formation of higher molecular weight derivatives), -C=C-; -C-C- bond cleavage, hydrogenation-dehydrogenation reactions, methane activation, and oxidation of organic solvents. These reactions are important in hydrocarbon refining, where an active catalyst is needed to improve the efficiency by which hydrogen is utilized. Malhotra et al. [20.89] inves tigated the properties of C60 and C70 as catalysts in the coupling, bond-cleavage, and dehydrogenation reactions of various organic solvents, such as 1,2-dinaphthylmethane (C21H16) and mesitylene [C6H3(CH3)3]. The results showed that C60 is relatively active in catalyzing these reactions. However, poor selectivity in the reaction products was observed, which implies that a wide range of products was obtained.

As an example of a hydrocarbon coupling reaction, fullerenes were extracted with organic solvents like mesitylene [C6H3(CH3)3] and 1,4-diethylbenzene and the products showed formation of dimers, trimers, and higher oligomers of these solvents, even at room temperature, while part of the C60 (10%) reacted with the solvent to form addition products.

As an example of a bond cleavage reaction, 1,2-dinaphthylmethane (C2iH16) was reacted in the presence of aromatic solvents and C60. The results showed that addition of C60 to the reaction medium increased the reaction rate constant by an amount that varied between 30% and 145%, depending on the aromatic solvent used for the reaction. In addition, C60 was also found to increase the fraction of dehydrogenated products from 1.3 to 33 [20.89].

Methane activation (formation of higher hydrocarbons from methane) using carbon soot containing fullerenes as a catalyst has also been investigated [20.88]. The results showed that the fullerene soot needed a lower onset temperature for the activation reaction compared to that of a high-surface-area activated carbon. However, it was not clear what role, if any, the nonfullerene components of the soot (e.g., carbon black) played in this reaction, since no reactions were carried out with a pure fullerene catalyst.

Fullerene-catalyzed oxidation and polymerization of organic solvents such as a-pinene (C10H16), 4-methyl-l-cyclohexene (C7H12), and pyridine (C5H5N) by C60 and C70 have been observed [20.90] in the presence of oxygen. Other forms of carbons (graphite and diamond) did not show any activity for the same oxidation reactions. It was found that only the solvents that possessed electron donor ability (such as the presence of carbon-carbon double bonds or a nitrogen lone pair) were converted by C60. Contrary to typical catalytic reactions in which the catalysts are not consumed, in these oxidation reactions spectrometric evidence showed that the fullerenes were consumed as they were broken down into fragments by the attack from hydrocarbon radicals formed during the reaction.

C60 has also been found to be a more active catalyst for the conversion of H2S into sulfur than neutral alumina, and charcoal mixed with alumina [20.92], This reaction constitutes the main process for the industrial production of sulfur. Proton NMR experiments on the products formed from the reaction of H2S bubbled through a solution of C60 in toluene showed evidence that C60 was also hydrogenated during this oxidation reaction.

Fullerenes can also be used as a catalyst support to increase the degree of dispersion of the active material. Nagashima et al. [20.93,94] synthesized a palladium-carrying polymer (see §10.10.2) in which the metal is highly dispersed in the C60 matrix so that its usage as a catalyst is enhanced. This C60-Pd polymer was found to catalyze the hydrogénation of diphenyl acetylene at standard temperature and pressure. Normally, a high loading of active metal is needed to catalyze such reactions.

20.3.4. Self-Assembled Monolayers

A rather popular research topic and possible future applications area for fullerenes concerns self-assembled monolayers [20.95,96] and a fullerene literature in this subject is now beginning to appear [20.97-100], We give here one simple example of a work already reported.

Layer-by-layer self-assembly has been demonstrated for a polyca-tion monolayer such as PAH [poly(allylamine hydrochloride)] or PPV [poly(phenylene vinylene)] followed by a polyanion monolayer such as sulfonated C60 [20.101]. For the sulfonated CM, several [-(CH2)4-S03H] sulfonic acid groups attach to each Qq molecule, and the layer-by-layer self-assembly is carried out by dipping a substrate into a solution of the polycation followed by dipping into the polyanion solution, and repeating this process, until the desired number of polycation-polyanion bilayers are assembled. The layer-by-layer assembly of bilayers is demonstrated in Fig. 20.18, where optical absorbance measurements at a wavelength of 420 nm are shown vs the number of C60-PAH bilayers. Since only the C60 layers are absorbing at this optical wavelength, Fig. 20.18 shows that the absorbance is proportional to the number of bilayers, which is independently measured by a profilometer [20.101]. In the case of the PPV-C60 bilayers, both the anion and cation layers are absorbing near the absorption peak (~420 nm), as shown in Fig. 20.19. In contrast, for the C60-PAH

Was this article helpful?

0 0

Post a comment