The preparation of metallofullerenes can be achieved in different ways. The methods most commonly used today are the arc-evaporation of impregnated graphitic rods and the laser-evaporation process (also called the laser-oven method). For the latter a rotating target of graphite and metal oxide (cemented with a pitch binder) is placed in an oven heated to about 1200 °C and irradiated with a doubled-frequency Nd: YAG laser under a stream of argon. The resulting fullerenes and metallofullerenes are swept along with the gas current and precipitated in the cooler zones at the end of the quartz tube. A temperature of not less than 800 °C is required for this process as neither empty nor filled fullerenes are formed below that value (refer to Section 2.3.4).

The arc-method uses the same procedure as successfully employed to produce empty fullerenes (refer to Section 2.3.3), with the sole difference that the anodes of graphite evaporated in a classical arc - apparatus are impregnated with metal oxides or carbides. (In the graphitic rods treated with metal oxides, the respective carbides will be generated, too, when heated to >1600 °C.) Besides empty species, the soot deposited on the cooler reactor walls contains various metallofullerenes.

Endofullerenes like [email protected] can further be obtained by ion implantation techniques. High - energy ions of the desired element are targeted at a thin film of fullerenes. However, the produced amounts are small and accordingly the analysis of products is complicated.

Apart from the expected [email protected] - the examination of soot containing metallofullerenes also reveals carbon cages of unusual sizes. [email protected] and [email protected] 2 are observed upon evaporation of a graphitic sample treated with La2 O3. The C8 2-metallofullerene exhibits the highest stability of all. It may, as an extract in toluene, even be stored under air without decomposing. [email protected], on the other hand, cannot be isolated due to its instability.

A similar picture arises with other elements: The C60-metallofullerenes are formed, yet they cannot be obtained in substance. Obviously, further factors influence stability in the generation of endofullerenes to the effect that smaller species

Table 2.8 Bonding energies AHB of metallofullerenes.


[email protected]

[email protected] [email protected]

[email protected]

AHb in kJmol-1


322 444


are less stable than bigger ones, with [email protected] being the most abundant. Just recently it has been possible, however, to isolate a few selected C60-metallofuller-enes, for example, [email protected] or [email protected]

In general, metallofullerenes are known for a multitude of elements by now, including alkali like Li or alkaline earth metals like Ca, Sr, and Ba, the elements of the scandium group Sc, Y, La and the lanthanides Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu as well as titanium, iron, and uranium. The elements of main group and subgroup 3 are much easier to encapsulate than for instance those of main group 2, which reflects in the bonding energies of individual metallofullerenes (relative to M + C82; Table 2.8). Endohedral compounds ofheterofuller-enes are know as well, it is possible, for instance, to enclose two lanthanum atoms in C79N.

This is indicative of a more general finding: In addition to simple endofullerenes [email protected] the encapsulation can yield species with several atoms confined to the same fullerene cage. The radius of the cavity inside C82 is 0.4 nm, which is large enough to host up to three atoms of rare earth metals. The generation of [email protected] becomes ever more probable with an increasing content of metal in the graphitic starting material. For scandium the complete series of [email protected], [email protected], [email protected] and, just recently, even [email protected] has been detected. Further on, [email protected] was isolated, which is one of the best studied dimetallofullerenes besides [email protected]

The formation of C80-metallofullerenes is a particularly interesting because C80 itself does not exist as an empty cage. The ground state of 8h-C80 exhibits an antiaromatic open- shell structure explaining this instability. (There are only two electrons situated in the quadruply degenerate HOMO.) In a respective endofuller-ene however, the electronic structure is effectively influenced by electron transfer from the metal atoms to the carbon cage. The latter turns into a stable closed-shell C|--ion by accepting six electrons. This may be effected, for example, by two atoms of lanthanum, resulting in the formation of [email protected] Today the metal atoms are known to circulate inside the icosahedral C80-cage. The movement could be shown by the line broadening in the 139La-NMR spectrum that is caused by the magnetic field resulting from this motion. The formation of C80-compounds is speculated to happen by the way of the so- called shrink- wrap mechanism that explains the generation of smaller species by an expulsion of C 8 -units from existing metallofullerenes and subsequent closure of the cage. This process may continue to cage sizes far smaller than C60 and stops only at about 44 (but not less than 36) carbon atoms. Further shrinking leads to destruction of the cage-like structure. The size limit is determined, among others, by the dimensions of the encapsulated metal atom.

It soon became evident from ESR-measurements that the metal atoms exist as positively charged ions inside the fullerenes. Obviously, they transfer their valence electrons to the cage. For instance [email protected] really is La3+ + @C8-. Yttrium expectedly behaves in an analogous manner, while scandium, due to its low- lying d - orbital, forms Sc2+ by transferring its 4s-electrons only. The tendency toward trivalent ions in Y and La is favored by their higher and more diffused d- orbitals. The charge status of lanthanide atoms bearing electrons in 4f-orbitals is subject to lively discussion. UV/Vis - spectra indicated that they are trivalent. Ce (4f- 5d16s2) and Gd (4f75d16s2) form trivalent ions accordingly, while elements such as Pr (4f36s2) and Nd (4f46s2) seem to transfer only two electrons. Eu (4f146s2) and Yb (4f76s2) each contribute their two 6s-electrons alone because a half or a fully occupied 4f-orbital is energetically favorable. Lutetium (4f145d16s2) has a configuration similar to scandium, yet it transmits the 5d- and one 6s- electron because 6s- electrons are stabilized by relativistic effects. Summing up, the circumstances are rather complicated by backbonding and by orbitals lying in close proximity.

A major issue arising after the discovery of metallofullerenes was whether the metal atoms were really contained inside the carbon cage because in principle, they could just as well be located on the fullerene ' s outer surface. There were several approaches to answer this question. Firstly, endofullerenes were demonstrated to expel C2-units in gas-phase fragmentation experiments (collision, laser impact, etc.), whereas exohedral compounds of the Fe^-J-type yield an intact buckminsterfullerene by cleaving the metal atom. Secondly, the endohedral structure could be shown by electron microscopic examination (HRTEM, STM) of the solid phase. Conclusive evidence, however, could only be adduced from studies employing synchrotron X-ray diffraction. In these studies, significant electron density was detected inside the fullerene, making clear at the same time that usually the metal atoms are not situated at the center of the carbon cage (Figure 2.46). This method could even provide information about the charge of the encapsulated metal. The scandium atom in [email protected] - for instance, was found not to be trivalent, but the structure rather corresponds to Sc2+ @C8-. The electronic properties of metallofullerenes are of special interest because they markedly deviate from those of empty fullerenes. The calculated ionization potentials and electron affinities for a series of metallofullerenes are collected in Table 2.9.

These values indicate that metallofullerenes can be both stronger electron donors and acceptors than their empty analogs C6 0 and C7 0. This observation is

Table 2.9 Electron affinities (EA) and ionization potentials (IP) of metallofullerenes.


[email protected]

[email protected]

[email protected]

[email protected]

[email protected]

[email protected]



EA in eV









IP in eV









See S. Nagase, K. Kobayashi, T. Akasaka, T. Wakahara, Endohedral Metallofullerenes: Theory, Electrochemistry, and Chemical Reactions, in: K. M. Kadish, R. S. Ruoff (editors), Fullerenes, Wiley Interscience, New York 2000.

See S. Nagase, K. Kobayashi, T. Akasaka, T. Wakahara, Endohedral Metallofullerenes: Theory, Electrochemistry, and Chemical Reactions, in: K. M. Kadish, R. S. Ruoff (editors), Fullerenes, Wiley Interscience, New York 2000.

Figure 2.46 Distribution of electron density in endofullerenes. It turns out that. for example, lanthanum exerts a motion within the cage (b) whereas scandium remains fixed at its position out of the cage's center (a) (© Wiley Interscience 2000).




——Jf- HOMO


f- 6s


ft 5d


[email protected]


Figure 2.47 Molecular orbital diagram of [email protected] (© ACS 1993).

confirmed by measurements of the reduction and oxidation potentials. For [email protected] C82, the species best characterized, a first oxidation step was found at +0.07 V, corresponding to a moderate electron donor (comparable to ferrocene) and explaining its stability to air. Furthermore there are five reduction potentials at -0.42 V, -1.37V, -1.53 V, -2.26V and -2.46V. Altogether, [email protected] is a better electron acceptor than an empty fullerene. As evident from Figure 2.47, the first reduction step above all is energetically favorable because the HOMO will be fully occupied by taking in an electron. The donor-acceptor characteristics of other metallofuller-enes may be deduced from the MO - diagram as well. They turn out to be good electron acceptors just as well, yet they can also dispose of unpaired electrons and be oxidized. The 4f-electrons in Ce -- Pr- - and Gd-fullerenes do not seem to be involved in this process.

Metallofullerenes that could actually be isolated have been examined for their reactivity as well. Exohedral functionalization was achieved, for example, by photochemical reactions, Diels-Alder addition, etc. In the reaction with disiliranes, the thermal variant was found to be viable besides the photochemical addition observed with empty fullerenes. This effect is caused by the better donor-acceptor properties. Contrasting the noble-gas fullerenes discussed in Section - the electron transfer from an inward metal atom to the carbon cage largely influences the chemical properties of the respective endofullerene. Dimetallofullerenes may also react with disiliranes in a thermal way, which again is explained by the acceptance of electrons to the fullerene cage. However, the number of metal atoms inside the cage is not crucial as long as the reduction potential is sufficiently low (in [email protected], for example, it is too high and the thermal reaction does not occur).

Was this article helpful?

0 0

Post a comment