We have performed both semiempirical as well as ab initio density functional theory calculations in order to investigate the structural stability of highly hydroxylated C60(OH)32 fullerenes, so-called fullerenols. Interestingly, we have found that low-energy atomic configurations are obtained when the OH groups are covering the C60 in the form of small hydroxyl islands. The previous formation of OH molecular domains on the carbon surface, stabilized by hydrogen bonds between neighboring OH groups, defines the existence of C60(OH)32 fullerene structures with some elongated C-C bonds, closed electronic shells, and large highest occupied-lowest unoccupied molecular orbital energy gaps, with the latter two being well-known indicators of high chemical stability in these kind of carbon compounds. The calculated optical absorption spectra show that the location of the first single dipole-allowed excitation strongly depends on the precise distribution of the OH groups on the surface, a result that, combined with optical spectroscopy experiments, might provide an efficient way to identify the structure of these kinds of fullerene derivatives. We found that the presence of a few coadsorbed oxygen species on the fullerene surface leads in general to the existence of C60(OH)32O(x) (x = 1-4) compounds in which some of the C-C bonds just below the O impurities are replaced by C-O-C bridge bonds, leading to the formation of stable carbon-opened structures in agreement with the recent experimental work of Xing et al. (J. Phys. Chem. B 2004, 108, 11473). Actually, a more dramatic cage destruction is obtained when considering multiply charged C60(OH)32O(x)(+/-m) (m = 2, 4, 6) species (that can exist in both gas-phase and aqueous environments), where now sizable holes made of 9- and 10-membered rings can exist in the carbon network. We believe that our results are important if the controlled opening of carbon cages is needed and it should be taken into account also in several technological applications where the permanent encapsulation of atomic or molecular species in these types of fullerene derivatives is required.

We have performed semiempirical modified-neglect-of-diatomic-overlap ͑to determine the most stable geometrical arrangements͒ as well as ab initio density-functional theory calculations ͑to obtain the electronic structure and total energies͒ at Tϭ0 to analyze the energetics and structural properties of OH species adsorbed on the external surfaces of spheroidal C 60 as well as on finite length open-ended armchair ͑6,6͒ and ͑8,8͒ carbon nanotubes. Interestingly we have found, for the low coverage regime, that the adsorbed OH groups prefer to organize as a hydroxyl cluster ͑having up to seven OH molecules͒ in only one side of the C 60 surface. The observed clustering leads to the formation of a new amphiphilic molecule that naturally explains the stability of C 60 (OH) n (nϳ9-12) Langmuir monolayers at the air-water interface observed by several authors, where it is thus only the highly hydroxylated part of the carbon cage the one that dissolves in water slightly. For 8 to 14 adsorbed OH groups, a second hydroxyl island is gradually stabilized on the opposite side of the carbon structure, and finally, with increasing coverage the coexistence of ringlike and cluster arrays of OH groups seems to lead to the complete solubility of the carbon compound. In all cases, the OH molecules have been found to occupy on-top sites with a C-O-H bond tilted away from the surface normal and no hydrogen bond formation between the adparticles is obtained, in contrast to what is normally observed in the compact OH hexagonal phases stabilized on extended metal substrates. The calculated vibrational frequencies of our adsorbates have a good correspondence with the experimental measurements and provide a clear signature of the clustering of the OH molecules at low coverages. Finally, OH adsorption on the external surface of cylindrical carbon structures leads to the formation of quasi-one-dimensional molecular arrays, consistent with the synthesis of metal nanowires on the surface of carbon nanotubes, their length being determined by the length of the nanotube.

No abstract