Carbon encapsulated nanoparticles: materials science and energy applications.

The technological implementation of electrochemical energy conversion and storage necessitates the acquisition of high-performance electrocatalysts and electrodes. Carbon encapsulated nanoparticles have emerged as an exciting option owing to their unique advantages that strike a high-level activity-stability balance. Ever-growing attention to this unique type of material is partly attributed to the straightforward rationale of carbonizing ubiquitous organic species under energetic conditions. In addition, on-demand precursors pave the way for not only introducing dopants and surface functional groups into the carbon shell but also generating diverse metal-based nanoparticle cores. By controlling the synthetic parameters, both the carbon shell and the metallic core are facilely engineered in terms of structure, composition, and dimensions. Apart from multiple easy-to-understand superiorities, such as improved agglomeration, corrosion, oxidation, and pulverization resistance and charge conduction, afforded by the carbon encapsulation, potential core-shell synergistic interactions lead to the fine-tuning of the electronic structures of both components. These features collectively contribute to the emerging energy applications of these nanostructures as novel electrocatalysts and electrodes. Thus, a systematic and comprehensive review is urgently needed to summarize recent advancements and stimulate further efforts in this rapidly evolving research field.

Fullerene Fragment Restructuring: How Spatial Proximity Shapes Defect-Rich Carbon Electrocatalysts.

Fullerene transformation emerges as a powerful route to construct defect-rich carbon electrocatalysts, but the carbon bond breakage and reformation that determine the defect states remain poorly understood. Here, we explicitly reveal that the spatial proximity of disintegrated fullerene imposes a crucial impact on the bond reformation and electrocatalytic properties. A counterintuitive hard-template strategy is adopted to enable the space-tuned fullerene restructuring by calcining impregnated C60 not only before but also after the removal of rigid silica spheres (∼300 nm). When confined in the SiO2 nanovoids, the adjacent C60 fragments form sp3 bonding with adverse electron transfer and active site exposure. In contrast, the unrestricted fragments without SiO2 confinement reconnect at the edges to form sp2-hybridized nanosheets while retaining high-density intrinsic defects. The optimized catalyst exhibits robust alkaline oxygen reduction performance with a half-wave potential of 0.82 V via the 4e- pathway. Copper poisoning affirms the intrinsic defects as the authentic active sites. Density functional theory calculations further substantiate that pentagons in the basal plane lead to localized structural distortion and thus exhibit significantly reduced energy barriers for the first O2 dissociation step. Such space-regulated fullerene restructuring is also verified by heating C60 crystals confined in gallium liquid and a quartz tube.

Two Birds with One Stone: Concurrent Ligand Removal and Carbon Encapsulation Decipher Thickness-Dependent Catalytic Activity.

A carbon shell encapsulating a transition metal-based core has emerged as an intriguing type of catalyst structure, but the effect of the shell thickness on the catalytic properties of the buried components is not well known. Here, we present a proof-of-concept study to reveal the thickness effect by carbonizing the isotropic and homogeneous oleylamine (OAm) ligands that cover colloidal MoS2. A thermal treatment turns OAm into a uniform carbon shell, while the size of MoS2 monolayers remains identical. When evaluated toward an acidic hydrogen evolution reaction, the calcined MoS2 catalysts deliver a volcano-type activity trend that depends on the calcination temperature. Rutherford backscattering spectrometry and depth-profiling X-ray photoelectron spectroscopy consistently provide an accurate quantification of the carbon shell thickness. The same variation pattern of catalytic activity and carbon shell thickness, aided by kinetic studies, is then persuasively justified by the respective limitations of electron and proton conductivities on the two branches of the volcano curve.

Laser-Triggered Bottom-Up Transcription of Chemical Information: Toward Patterned Graphene/MoS2 Heterostructures.

Efficiently assembling heterostructures with desired interface properties, stability, and facile patternability is challenging yet crucial to modern device fabrication. Here, we demonstrate an interface coupling concept to bottom-up construct covalently linked graphene/MoS2 heterostructures in a spatially defined manner. The covalent heterostructure domains are selectively created in analogy to the traditional printmaking technique, enabling graphic patterns at the bottom MoS2 layer to be precisely transferred to the top graphene layer. This bottom-up connection and transcription of chemical information is achieved simply via laser beam irradiation. Our approach opens up a new paradigm for heterostructure construction and integration. It enables the efficient generation and real-time visualization of spatially well-resolved covalent graphene/MoS2 heterostructures, facilitating further design and integration of patterned heterostructures into new generations of high-performance devices.

Covalent 2D Patterning, Local Electronic Structure and Polarization Switching of Graphene at the Nanometer Level.

A very facile and efficient protocol for the covalent patterning and properties tuning of graphene is reported. Highly reactive fluorine radicals were added to confined regions of graphene directed by laser writing on graphene coated with 1-fluoro-3,3-dimethylbenziodoxole. This process allows for the realization of exquisite patterns on graphene with resolutions down to 200 nm. The degree of functionalization, ranging from the unfunctionalized graphene to extremely high functionalized graphene, can be precisely tuned by controlling the laser irradiation time. Subsequent substitution of the initially patterned fluorine atoms afforded an unprecedented graphene nanostructure bearing thiophene groups. This substitution led to a complete switch of both the electronic structure and the polarization within the patterned graphene regions. This approach paves the way towards the precise modulation of the structure and properties of nanostructured graphene.

Covalently Doped Graphene Superlattices: Spatially Resolved Supratopic- and Janus-Binding.

Rational design and fabrication of graphene nanoarchitectures with multifunctionality and multidimensionality remains quite a challenge. Here, we present a synthetic sequence, based on the combination of two advanced patterned-functionalization principles, namely, laser-writing and poly(methyl methacrylate) (PMMA)-assisted lithographic processes, leading to unprecedented covalently doped graphene superlattices. Spatially resolved supratopic- and Janus-binding were periodically weaved on the graphene sheet, leading to four different types of zones with distinct chemical doping and structural properties. Notably, this is also the first realization of patterned Janus graphene. The elaborate chemical doping with micrometer resolution is unequivocally evidenced by scanning Raman spectroscopy (SRS) and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS). The design of the pattern as well as the degree of chemical doping on both opposite sides of graphene can be easily manipulated, rendering exciting potential for graphene nanosystems.

Crystallographic Characterization of Ti2C2@D3h(5)-C78, Ti2C2@C3v(8)-C82, and Ti2C2@Cs(6)-C82: Identification of Unsupported Ti2C2 Cluster with Cage-Dependent Configurations.

Fullerene cages are ideal hosts to encapsulate otherwise unstable metallic clusters to form endohedral metallofullerenes (EMFs). Herein, a novel Ti2C2 cluster with two titanium atoms bridged by a C2-unit has been stabilized by three different fullerene cages to form Ti2C2@D3h(5)-C78, Ti2C2@C3v(8)-C82, and Ti2C2@Cs(6)-C82, representing the first examples of unsupported titanium carbide clusters. Crystallographic results show that the configuration of the Ti2C2 cluster changes upon cage variation. In detail, the Ti2C2 cluster adopts a butterfly shape in Ti2C2@C3v(8)-C82 and Ti2C2@Cs(6)-C82 with Ti-C2-Ti dihedral angles of 156.35 and 147.52° and Ti-Ti distances of 3.633 and 3.860 Å, respectively. In sharp contrast, a stretched planar geometry of Ti2C2 is observed in Ti2C2@D3h(5)-C78, where a Ti-C2-Ti angle of 176.87° and a long Ti-Ti distance of 5.000 Å are presented. Consistently, theoretical calculations reveal that the cluster configuration is very sensitive to the cage shape which eventually determines the electronic structures of the hybrid EMF-molecules, thus adding new insights into modern coordination chemistry.

Spatially Resolved Bottom-Side Fluorination of Graphene by Two-Dimensional Substrate Patterning.

Patterned functionalization can, on the one hand, open the band gap of graphene and, on the other hand, program demanding designs on graphene. The functionalization technique is essential for graphene-based nanoarchitectures. A new and highly efficient method was applied to obtain patterned functionalization on graphene by mild fluorination with spatially arranged AgF arrays on the structured substrate. Scanning Raman spectroscopy (SRS) and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS) were used to characterize the functionalized materials. For the first time, chemical patterning on the bottom side of graphene was realized. The chemical nature of the patterned functionalization was determined to be the ditopic scenario with fluorine atoms occupying the bottom side and moieties, such as oxygen-containing groups or hydrogen atoms, binding on the top side, which provides information about the mechanism of the fluorination process. Our strategy can be conceptually extended to pattern other functionalities by using other reactants. Bottom-side patterned functionalization enables utilization of the top side of a material, thereby opening up the possibilities for applications in graphene-based devices.

Preferential Formation of Mono-Metallofullerenes Governed by the Encapsulation Energy of the Metal Elements: A Case Study on Eu@C2n (2n=74-84) Revealing a General Rule.

Encapsulating one to three metal atoms or a metallic cluster inside fullerene cages affords endohedral metallofullerenes (EMFs) classified as mono-, di-, tri-, and cluster-EMFs, respectively. Although the coexistence of various EMF species in soot is common for rare-earth metals, we herein report that europium tends to prefer the formation of mono-EMFs. Mass spectroscopy reveals that mono-EMFs (Eu@C2n ) prevail in the Eu-containing soot. Theoretical calculations demonstrate that the encapsulation energy of the endohedral metal accounts for the selective formation of mono-EMFs and rationalize similar observations for EMFs containing other metals like Ca, Sr, Ba, or Yb. Consistently, all isolated Eu-EMFs are mono-EMFs, including Eu@D3h (1)-C74 , Eu@C2v (19138)-C76 , Eu@C2v (3)-C78 , Eu@C2v (3)-C80 , and Eu@D3d (19)-C84 , which are identified by crystallography. Remarkably, Eu@C2v (19138)-C76 represents the first Eu-containing EMF with a cage that violates the isolated-pentagon-rule, and Eu@C2v (3)-C78 is the first C78 -based EMF stabilized by merely one metal atom.

Bonding inside and outside Fullerene Cages.

Concrete crystallographic results of endohedral metallofullerenes (EMFs) disclose that the bonding within the metallic clusters and the interactions between the metal ions and the cage carbon atoms, which are closely associated with the coordination ability of the metal ions, play essential roles in determining the stability, the molecular structure, and the chemical behavior of the hybrid EMF molecules, in addition to the previously recognized charge transfer from metal to cage. For the carbide cluster metallofullerenes, a "size effect" between the encapsulated metallic cluster and the fullerene cage has been suggested. Thus, through the geometric effect, a series of giant fullerenes (C90-C104) have been stabilized by encapsulating a large La2C2 cluster, which adopts different configurations in accordance with cage size and shape. Interestingly, the crystallographic analysis of La2C2@ D5(450)-C100 has led to the direct observation of the axial compression of short carbon nanotubes caused by the internal stress. Additionally, the defective C2(816)-C104 cage is viewed to be a precursor that can transform into the other three ideal tubular fullerene cages, presenting crystallographic evidence for the top-down formation mechanism of fullerenes. Structural characterization of Y2C2@C108 confirms a linear carbide cluster inside the large cage, indicative of a geometric effect of cage size on the bonding behavior of the internal cluster. Apart from the carbide realm, direct metal-metal bonding is observed between the two seemingly repulsive Lu2+ ions in Lu2@C82-86, adding new insights into current coordination chemistry. Meanwhile, the bonding state between the metal ions inside the cage (e.g., in La2@ I h(7)-C80) and even the configuration of the internal metallic cluster (e.g., in Sc3C2@ I h(7)-C80) can be readily controlled by exohedral radical addition, illuminating their future applications as single molecule magnets and in electronics. In addition, observation of the unexpected dimerization between two paramagnetic Y@ Cs(6)-C82 molecules suggests a spin-induced bonding behavior, which depends closely on the cage geometry. In contrast, synergistic effect of both electronic and geometric parameters has led to the formation of the unprecedented [6,6,6]-Lewis acid-base adduct of Sc3N@ I h(7)-C80. However, introduction of an oxygen atom gives rise to the corresponding normal carbene adducts for both Sc3N@ I h(7)-C80 and Lu3N@ I h(7)-C80, presenting an unexpected way of steric hindrance release. Remarkably, the Lewis acid-base complexation is demonstrated to be a facile way toward isomerically pure metallofullerene derivatives with surprisingly high regioselectivity and quantitative conversion yield for Sc2C2@ C3 v(8)-C82. This Account aims to give an advanced summary of the recent achievements in research of EMFs, focusing mainly on the interplay between the internal metallic species and the surrounding cages through bond formation or cleavage. Perspectives suggesting the future developments of EMFs are also given in the last section.

Reactions between N-Heterocyclic Carbene and Lutetium-Metallofullerenes: High Regioselectivity Directed by Electronic Effect in Addition to Steric Hindrance.

The Lewis acid-base pairing reaction between strained N-heterocyclic carbene (NHC) and endohedral metallofullerenes (EMFs) is an efficient strategy to obtain stable derivatives in a highly regioselective manner. Herein, we report an in-depth study on the reactions between 3-dimesityl-1 H-imidazol-3-ium-2-ide (1) and three different EMFs, namely, Lu3N@ Ih(7)-C80, Lu2@ C3 v(8)-C82, and Lu2@ C2 v(9)-C82, respectively. Only one monoadduct is obtained for each EMF under certain conditions, demonstrating surprisingly high regioselectivity and exclusive formation of monoadducts. X-ray results of the derivatives of Lu3N@C80 reveal that an epoxide adduct (2a) with a specific [6,6,6]-carbon atom of the C80 cage singly bonded to the normal carbene center (C2) of the NHC is obtained under ambient condition, whereas a pure argon atmosphere gives 2b with an abnormal C5-bonding structure. In contrast, the derivatives of Lu2@C82 (3 and 4) are both normal C2-bonding [5,6,6]-adducts without oxygen addition, even though air is involved in the reaction. Our theoretical results confirm that the remarkably high regioselectivity and the quantitative formation of monoadducts are direct result from the distributions of molecular orbital and electrostatic potential on the cage surfaces in addition to the previously assumed steric hindrance between the fullerene cage and the NHC moiety.

High-Nuclearity Heterometallic tert-Butylethynide Clusters Assembled with tert-Butylphosphonate.

tert-Butylphosphonic acid and lanthanide precursors were employed to construct two high-nuclearity hybrid silver(I)-ytterbium(III) phosphonate clusters: compound 1 consists of a Ag16 ethynide cluster fused with a trinuclear hydroxoytterbium phosphonate cluster, whereas compound 2 is composed of two Ag16 ethynide clusters bridged by a hexanuclear oxo/hydroxoytterbium phosphonate cluster. Using transition-metal-substituted lacunary polyoxotungstates in place of the lanthanide reactant, new phosphonate-functionalized silver(I)-copper(II) ethynide clusters [Ag34 Cu6 (3) and Ag37 Cu6 (4)] and silver(I) ethynide clusters [Ag51 (5) and Ag72 (6)] were obtained. The structures of complexes 3-6 feature core-shell arrangements, in which silver(I)-copper(II) or silver(I) ethynide cluster shells stabilized by peripheral phosphonate ligands enclose different kinds of tungstate core templates.

Isolation and Crystallographic Characterization of Lu3 N@C2n (2n=80-88): Cage Selection by Cluster Size.

The small Sc3 N cluster has only been found in such small cages as C2n (2n=68, 78, 80, 82), whereas the large M3 N (M=Y, Gd, Tb, Tm) clusters choose those larger cages C2n (2n=82-88). Herein, concrete experimental evidence is presented to establish the size effect of the internal metallic cluster on selecting the outer cage of endohedral metallofullerenes (EMFs) by using a medium-sized metal, lutetium, which possesses an ionic radius between Sc and Gd. A series of lutetium-containing EMFs have been obtained and their structures are unambiguously determined as Lu3 N@Ih (7)-C80 , Lu3 N@D5h (6)-C80 , Lu3 N@C2v (9)-C82 , Lu3 N@Cs (51365)-C84 , Lu3 N@D3 (17)-C86 , and Lu3 N@D2 (35)-C88 by single-crystal X-ray diffraction crystallography. It was confirmed that the encaged Lu3 N cluster always adopts a planar geometry in Lu3 N@C80-88 isomers to ensure substantial metal-cage/metal-nitrogen interactions. As a result, the Lu3 N cluster selects the C2v (9)-C82 cage, which also encapsulates Sc3 N, instead of the Cs (39663)-C82 cage which is more suitable for M3 N (M=Y, Gd, Tb, Tm). However, different from Sc3 N, Lu3 N can also template the C84-88 cages which are absent for Sc3 N-containing EMFs, confirming clearly the size effect of the internal cluster on selecting the outer cage.

Understanding Charge-Transfer Characteristics in Crystalline Nanosheets of Fullerene/(Metallo)porphyrin Cocrystals.

Cocrystals in the form of crystalline nanosheets comprised of C70 and (metallo)porphyrins were prepared by using the liquid-liquid interfacial precipitation (LLIP) method where full control over the morphologies in the C70/(metallo)porphyrins nanosheets has been accomplished by changing the solvent and the relative molar ratio of fullerene to (metallo)porphyrin. Importantly, the synergy of integrating C70 and (metallo)porphyrins as electron acceptors and donors, respectively, into nanosheets is substantiated in the form of a near-infrared charge-transfer absorption. The presence of the latter, as reflection of ground-state electron donor-acceptor interactions in the nanosheets, in which a sizable redistribution of charge density from the electron-donating (metallo)porphyrins to the electron-accepting C70 occurs, leads to a quantitative quenching of the localized (metallo)porphyrin fluorescence. Going beyond the ground-state characterization, excited-state electron donor-acceptor interactions are the preclusion to a full charge transfer featuring formation of a radical ion pair state, that is, the one-electron reduced fullerene and the one-electron oxidized (metallo)porphyrin.

Stabilization of Giant Fullerenes C2(41)-C90, D3(85)-C92, C1(132)-C94, C2(157)-C96, and C1(175)-C98 by Encapsulation of a Large La2C2 Cluster: The Importance of Cluster-Cage Matching.

Successful isolation and unambiguous crystallographic assignment of a series of higher carbide cluster metallofullerenes present new insights into the molecular structures and cluster-cage interactions of endohedral metallofullerenes. These new species are identified as La2C2@C2(41)-C90, La2C2@D3(85)-C92, La2C2@C1(132)-C94, La2C2@C2(157)-C96, and La2C2@C1(175)-C98. This is the first report for these new cage structures except for D3(85)-C92. Our experimental and theoretical results demonstrate that La2C92-106 are more inclined to exist stably in the carbide form La2C2@C90-104 rather than as the dimetallofullerenes La2@C92-106, which are rationalized by considering a synergistic effect of inserting a C2 unit into the cage, which ensures strong metal-cage interactions by partially neutralizing the charges from the metal ions and by fulfilling the coordination requirement of the La3+ ions as much as possible.

B-N-B Bond Embedded Phenalenyl and Its Anions.

Despite the great potential of heteroatom-containing polycyclic aromatic hydrocarbons in organic optoelectronics, there are very limited reports on heteroaromatics containing a B-N-B bond in the π-scaffold. Herein, stable 1,9-dibora-9a-azaphenalenyl (DBAP) derivatives, named BNB-embedded phenalenyls, are presented. The DBAP skeleton contains a three-center two-π-electron B-N-B moiety with 12 π-electrons and can be regarded as the isoelectronic structure of the phenalenyl cation. Chemical reduction of the phenyl derivative of DBAP by potassium generated the dianion containing 14 delocalized π-electrons, which can be regarded as the isoelectronic structure of the phenalenyl anion. The dianion is sandwiched and stabilized by two bulky [K([18]crown-6)]+ counterions according to its X-ray structure. However, its monoanion (an isoelectronic structure of the henalenyl radical) generated by mixing equal moles of neutral compound and dianion gave an unusual B-N-B-embedded benzo[cd]fluoranthene dianion, which again was confirmed by X-ray crystallographic analysis. The new dianion containing 20 π-electrons is highly aromatic and is further stabilized by [K([18]crown-6)]+ counterions. An aromaticity driven rearrangement mechanism was proposed for this unusual transformation. Based on X-ray structures and theoretical calculations, the B-N-B moiety in the neutral and anionic DBAP participates in the π-electron delocalization along the whole DBAP skeleton like their phenalenyl cation/radical/anion counterparts, but with more localized character. Therefore, our studies report the first synthesis and characterization of a B-N-B-embedded phenalenyl and its anionic species, which show unique electronic structure and unusual reactivity different from that of their all-carbon phenalenyl analogues.

Lu2@C2n (2n = 82, 84, 86): Crystallographic Evidence of Direct Lu-Lu Bonding between Two Divalent Lutetium Ions Inside Fullerene Cages.

Although most of the M2C2n-type metallofullerenes (EMFs) tend to form carbide cluster EMFs, we report herein that Lu-containing EMFs Lu2C2n (2n = 82, 84, 86) are actually dimetallofullerenes (di-EMFs), namely, Lu2@Cs(6)-C82, Lu2@C3v(8)-C82, Lu2@D2d(23)-C84, and Lu2@C2v(9)-C86. Unambiguous X-ray results demonstrate the formation of a Lu-Lu single bond between two lutetium ions which transfers four electrons in total to the fullerene cages, thus resulting in a formal divalent state for each Lu ion. Population analysis indicates that each Lu atom formally donates a 5d electron and a 6s electron to the cage with the remaining 6s electron shared with the other Lu atom to form a Lu-Lu single bond so that only four electrons are transferred to the fullerene cages with the formal divalent valence for each lutetium ion. Accordingly, we confirmed both experimentally and theoretically that the dominating formation of di-EMFs is thermodynamically very favorable for Lu2C2n isomers.

Evidence of Oxygen Activation in the Reaction between an N-Heterocyclic Carbene and M3N@Ih(7)-C80: An Unexpected Method of Steric Hindrance Release.

We herein demonstrate for the first time the unexpected oxygen-involving reaction between M3N@Ih(7)-C80 (M = Sc, Lu) and 1,3-bis(diisopropylphenyl)imidazol-2-ylene (1). By introducing a tiny amount of oxygen into the reaction, unprecedented products (2a for Sc3N@C80 and 3a for Lu3N@C80) with the normal carbene center C2 singly bonded to a triple hexagonal junction (THJ) cage carbon together with an oxygen atom bridging the same THJ carbon atom and a neighboring carbon atom forming an epoxy structure are obtained. In situ mechanism study, in combination with theoretical calculations, reveals that the bond-breaking peroxidation facilitates the formation of the unexpected products 2a and 3a, providing new insight into fullerene chemistry.

Lewis Acid-Base Adducts of Sc2C2@C3v(8)-C82/N-Heterocyclic Carbene: Toward Isomerically Pure Metallofullerene Derivatives.

The addition of a bulky N-heterocyclic carbene (NHC) to Sc2C2@C3v(8)-C82 affords two monoadducts (2a and 2b) quantitatively and regioselectively, representing the first examples of Lewis acid-base pairs of metal carbide cluster fullerenes. 2b is likely a kinetically favorable labile product that cannot be isolated from the solution. The crystallographic results of 2a unambiguously demonstrate that one polarized C-C single bond is formed between the normal carbene site C2N of the NHC and a specific [5,6,6]-carbon atom out of 17 types of nonequivalent cage carbon atoms of Sc2C2@C3v(8)-C82. Theoretical calculations demonstrate that the high regioselectivity, the unexpected addition pattern, and the quantitative formation of monoadducts are synergistic results from the cage geometry and electron distribution on the cage.

Isolation and Crystallographic Characterization of La2C2@Cs(574)-C102 and La2C2@C2(816)-C104: Evidence for the Top-Down Formation Mechanism of Fullerenes.

Tubular higher fullerenes are prototypes of finite-length end-capped carbon nanotubes (CNTs) whose structures can be accurately characterized by single-crystal X-ray diffraction crystallography. We present here the isolation and crystallographic characterization of two unprecedented higher fullerenes stabilized by the encapsulation of a La2C2 cluster, namely, La2C2@Cs(574)-C102, which has a perfect tubular cage corresponding to a short (10, 0) zigzag carbon nanotube, and La2C2@C2(816)-C104 which has a defective cage with a pyracylene motif inserting into the cage waist. Both cages provide sufficient spaces for the large La2C2 cluster to adopt a stretched and nearly planar configuration, departing from the common butterfly-like configuration which has been frequently observed in midsized carbide metallofullerenes (e.g., Sc2C2@C80-84), to achieve strong metal-cage interactions. More meaningfully, our crystallographic results demonstrate that the defective cage of C2(816)-C104 is a starting point to form the other three tubular cages known so far, i.e., D5(450)-C100, Cs(574)-C102, and D3d(822)-C104, presenting evidence for the top-down formation mechanism of fullerenes. The fact that only the large La2C2 cluster has been found in giant fullerene cages (C>100) and the small clusters M2C2 (M = Sc, Y, Er, etc.) are present in midsized fullerenes (C80-C86) indicates that geometrical matching between the cluster and the cage, which ensures strong metal-cage interactions, is an important factor controlling the stability of the resultant metallofullerenes, in addition to charge transfer.