Endohedral Metallofullerenes: Fullerenes with Metal Inside presents a comprehensive survey of the current state of knowledge on endohedral metallofullerenes, from preparation to functionalization, reactivity and applications. Following a brief historical overview, the book describes methods for synthesis, extraction, separation and purification, and provides an insight into the molecular and crystal structures. Subsequent chapters discuss various categories of endohedral metallofullerenes based on the encapsulated species, including carbides, nitrides, sulphides, oxides, non-metal and non-IPR endohedral metallofullerenes, followed by scanning tunneling microscopy studies and the examination of electronic, vibrational, magnetic and optical properties. The book concludes with chapters addressing the chemical functionalization of endohedral metallofullerenes, and applications ranging from solar cells to biomedicine.
1.1 THE FIRST EXPERIMENTAL EVIDENCE OF METALLOFULLERENES
Endohedral metallofullerenes, the fullerenes with metal atom(s) inside, are an interesting class of fullerene materials because electron transfer from the encaged metal atom to the carbon cage has been known to occur and this dramatically alters the electronic and magnetic properties of the fullerenes.
Just a week after the first experimental observation of the "magic number" soccerball-shaped C60 in a laser-vaporized cluster beam mass spectrum by Kroto et al. , the same research group of Kroto, Smalley, and co-workers also found a magic number feature due to LaC60 in a mass spectrum prepared by laser vaporization of a LaCl3 impregnated graphite rod . They observed a series of Cn+ and LaCn+ ion species with LaC60+ as a magic number ion in the mass spectrum (Figure 1.1) and concluded that a La atom was encaged within the (then hypothetical) soccerball-shaped C60. This was obviously the first proposal of the so-called "endohedral metallofullerene" concept based on experiments. They first tried Fe with no success and found that La is a "correct" atom for encapsulation within fullerenes. It is interesting to note that even today Fe has not been encapsulated by fullerenes.
Figure 1.1 Laser-vaporization supersonic cluster-beam time-of-flight mass spectrum of various lathanum-carbon clusters. LaC60 is seen as an enhanced (magic number) peak.
Reprinted with permission from Ref. . Copyright 1985 American Chemical Society
Further circumstantial (not direct) evidence that metal atoms may be encaged in C60 was also reported by the Rice group, showing that LaC60+ ions did not react with H2, O2, NO, and NH3 . This suggests that reactive metal atoms are protected from the surrounding gases and are indeed trapped inside the C60 cage.
In fact, the first direct evidence of the soccerball (truncated icosahedron) C60 was amply demonstrated in 1990 by a historical experiment done by Kräetschmer, Huffman, and co-workers. They succeeded in producing macroscopic quantities of soccerball-shaped C60 by using resistive heating of graphite rods under a He atmosphere [4,5]. The resistive heating method was then superseded by the so-called contact arc discharge method  since the arc discharge method can produce fullerenes order of magnitudes larger than that by resistive heating. Since then, the arc discharge method has become a standard method for fullerene synthesis.
The first production of macroscopic quantities of endohedral metallofullerenes was also reported by Smalley and co-workers . They used high-temperature laser vaporization of La2O3/graphite composite rods together with the corresponding contact arc technique to produce various sizes of La-containing metallofullerenes. Contrary to the previous expectation, only the La@C82 fullerene was extracted by toluene solvent even though La@C60 and La@C70 were also seen in the mass spectra of the sublimed film from soot. In other words, the major La-metallofullerene with air and solvent stability is La@C82, and La@C60 and La@C70 are somehow unstable in air and solvents.
Figure 1.2 shows a Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrum of hot toluene extract of fullerene materials produced by laser vaporization of a 10% La2O3/graphite composite rod . In addition to empty fullerenes, only the La@C82 metallofullerene is seen and La@C60 and La@C70 are completely absent in the mass spectrum of the solvent extracts. The special behavior of the La@C82 fullerene was soon confirmed by Whetten and co-workers . However, they also observed that at relatively high loading ratios of La2O3 in composite rods a di-lanthanofullerene, La2@C80, was also produced by the resistive-heating method and found to be another solvent-extractable major lanthanofullerene [8,9].
Figure 1.2 An FT-ICR mass spectrum of hot toluene extract of fullerene soot produced by high-temperature laser vaporization of a 10% La2O3/graphite composite rod.
Reproduced with permission from Ref. . Copyright 1991 American Chemical Society
1.2 EARLY YEARS OF METALLOFULLERENE RESEARCH
The first important information on the electronic structure of La@C82 was provided by the IBM Almaden research group. The charge state of the encaged La atom was studied by Johnson et al.  using electron spin resonance (ESR). The ESR hyperfine splitting (hfs) analysis of La@C82 revealed that the La atom is in a trivalent +3 charge state and that the formal charge state of La@C82 should be written as La3+@C823-: three outer electrons of La transferring to the C82 cage .
Several other research groups extended their work to endohedral yttrium compounds. The Rice-Minnesota University  and Nagoya University  research groups also reported solvent-extractable Y@C82 and Y2@C82 fullerenes and observed the ESR hfs of Y@C82. From the hfs analyses both groups concluded that the charge state of the Y atom is also +3 and that a similar intra-fullerene electron transfer was taking place in Y@C82 as La@C82. These results were also confirmed by the NRL group . In addition, they also reported the production of mixed di-metallofullerenes such as (LaY)@C80. McElvany  reported the production of a series of yttrium fullerenes, Ym@Cn including Y@C82, by direct laser vaporization of samples containing graphite, yttrium oxide, and fullerenes in the gas phase.
Scandium metallofullerenes were also produced in macroscopic quantity and solvent-extracted by Shinohara et al.  and Yannoni et al. . Interestingly, the Sc fullerenes exist in extracts as a variety of species (mono-, di-, tri-, and even tetra-scandium fullerenes), typically as Sc@C82, Sc2@C74, Sc2@C82, Sc2@C84, Sc3@C82, and Sc4@C82. It was found that Sc3@C82 was also an ESR-active species whereas di- and tetra-scandium fullerenes such as Sc2@C84 and Sc4@C82 were ESR-silent. (See Section 3.2.2 for the present correct assignment for some of the di- and tri-scandium metallofullerenes.) A detailed discussion on the electronic structures of the scandium fullerenes accrued from these ESR experiments is given in Sections 5.1 and 5.2.
The formation of lanthanide metallofullerenes R@C82 (R = Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Yb, and Lu) was also reported by the UCLA  and SRI international  groups. These metallofullerenes were also found to be based on the C82 fullerene.
In addition to group 3 (Sc, Y, La) and the lanthanide metallofullerenes, group 2 metal atoms (Ca, Sr, Ba) also form endohedral metallofullerenes, and have been produced and isolated in milligram quantity [20-26]. These metal atoms have been encaged not only by C82 and C84 but also by smaller fullerenes such as C72, C74, and C80. Furthermore, group 4 metallofullerenes (Ti, Zr, Hf) were synthesized and isolated [27,28].
C60-based endohedral fullerenes which were produced but not isolated in the early days are Ca@C60 [29-33] and U@C60 . The Ca@C60 and U@C60 fullerenes are unique metallofullerenes in which Ca and U atoms are encaged by C60, and are quite different from group 3 and lanthanide, R@C82 type, metallofullerenes. An ab initio self-consistent field (SCF) Hartree-Fock calculation indicates that the Ca ion in Ca@C60 is displaced by 0.7 Å from the center and that the electronic charge of Ca is 2+ [30,34]. A similar theoretical prediction has been made on Sc@C60 by Scuseria and co-workers . Metallofullerenes based on C60 are known to be unstable in air and in normal fullerene solvents such as toluene and carbon disulfide. We will discuss the stability and properties of C60-based metallofullerenes together with an inability to extract and purify these in Chapter 14. The metal atoms which have been reported to form endohedral metallofullerenes are shown in Table 1.1.
Table 1.1 The "bucky periodic table" showing the elements which have been reported to form endohedral metallofullerenes and isolated as purified forms. The black elements form endohedral metallofullerenes which have been purified, whereas the gray elements form endohedral non-metallofullerenes
1.3 CONVENTIONAL AND IUPAC NOMENCLATURE FOR METALLOFULLERENES
The symbol @ is conventionally used to indicate that atoms listed to the left of the @ symbol are encaged in the fullerenes. For example, a C60-encaged metal species (M) is then written as M@C60 . The corresponding IUPAC nomenclature is different from this conventional M@C60 representation. It is recommended by IUPAC that La@C82 should be called  fullerene-incar-lanthanum and be written as iLaC82 . However, throughout this book the conventional M@C2n form is used for endohedral metallofullerenes for brevity, unless otherwise noted.
-  Kroto H W, Heath J R, O'Brien S C et al. 1985 C60: Buckminsterfullerene Nature 318 162
-  Heath J R, O'Brien S C, Zhang Q et al. 1985 Lanthanum complexes of spheroidal carbon shells J. Am....