Crystallographic characterization of Y₂C_2n (2n = 82, 88–94): direct Y–Y bonding and cage-dependent cluster evolution

Abstract

Direct yttrium–yttrium bonding has been a long-sought puzzle in organometallic chemistry to understand the catalytic processes that involve yttrium. Herein, we report the first crystallographic authentication of direct Y–Y bonding inside the hollow cavity of fullerene cages by forming endohedral metallofullerenes (EMFs). Based on an efficient separation/purification process, which involves Lewis-acid treatment and HPLC separation, we have obtained sufficient amounts of a series of Y₂C_2n (2n = 82, 88–94) isomers for systematic studies. The unambiguous single-crystal X-ray diffraction (XRD) crystallographic results show that two of them are di-EMFs, namely Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂, in which the long-sought Y–Y single bond between the two divalent yttrium ions is experimentally confirmed for the first time. In contrast, all the other EMFs with relatively large cages are carbide cluster metallofullerenes (CCMFs), namely, Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀ and Y₂C₂@C₂(61)-C₉₂. Consistently, our computational results prove that these experimentally obtained EMFs are all abundant at the high temperatures for fullerene formation (∼1500–3000 K) due to the strong coordination ability of yttrium ions, which enables the formation of either direct Y–Y bonds (for Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂) or the inclusion of a C₂-unit (in Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀ and Y₂C₂@C₂(61)-C₉₂). Our results suggest that metal atoms such as yttrium tend to adopt a low valence state during the arc-discharge process because of the presence of the highly reductive carbon plasma in the chamber, enabling the formation of an Y₂ dimer with direct Y–Y bonding in small cages like C₈₂.

---

Introduction

The pursuit of novel metal–metal bonds, either supported or unsupported, is of fundamental importance for the in-depth understanding of industrial catalysis, surface chemistry and even bio-inorganic processes.¹ Unsupported bonds are more attractive because ligand bridging may alter the nature of metal–metal bonding.² Metal–metal bonds involving elements in the f-block, especially rare-earth metals including scandium and yttrium, have attracted much attention to advance new bonding theory or novel chemical transformations.³ Most of the studies have focused on the synthesis and characterization of such complexes containing metal–metal bonds between a rare-earth metal and a main group or a transition metal. To the best of our knowledge, confirmed direct metal–metal bonding between pure rare-earth elements is very rare.²

The advent of fullerenes whose hollow cavity can host a variety of metallic elements to form endohedral metallofullerenes (EMFs) presents a huge possibility for the investigation of metal–metal interactions by using single crystal X-ray diffraction (XRD) studies.^4–6 Traditional EMFs are those encapsulating one or two metal atoms, which transfer a certain number of electrons to the cage.^7,8 When multiple metal ions are introduced into the cages, a non-metal element is generally required to stabilize the whole cluster to form species such as a metal carbide,^9–11 nitride,^12,13 sulphide,¹⁴ oxide^15–17 and even cyanide.^18,19

In principle, the positively charged metal ions suffer from strong Coulomb repulsion inside fullerene cages. For example, the estimated repulsion energy between the two La ions in La₂@C₈₀ is as high as 10 eV, which is comparable to the dissociation energy of the strongest covalent bonds like the N–N triple bond (9.8 eV).²⁰ Accordingly, direct (unsupported) metal–metal bonding has not been realized in EMFs until the theoretical analysis of Y₂@C₇₉N revealing a single-electron bond between the two repulsive metal ions.²¹ In the following studies, single-electron bonds between two metal ions have also been observed for functionalized or reduced M₂@C₈₀ compounds (M = La, Y and Dy).^22–24 These results are explained theoretically by Popov and co-workers by considering metal-localized HOMOs which exhibit pronounced metal–metal bonding character.²⁰ In addition, theoretical calculations of Lu₂@C₇₆ (ref. 25) predicted the presence of a normal metal–metal bond between the two Lu²⁺ ions, and crystallographic evidence for direct Lu–Lu bonding has been reported recently by our group in a series of Lu-containing EMFs, such as Lu₂@C_2n.^26,27 These results suggest that the confined inner space of fullerene cages can restrict the separation of the repulsive metal ions so as to facilitate the formation of metal–metal bonds, which stimulates us to seek for other unsupported metal–metal bonds between rare earth metals.

Herein, we report the isolation and systematic characterization of a series of di-yttrium EMFs, namely Y₂@C_s(6)-C₈₂, Y₂@C_3v(8)-C₈₂, Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀ and Y₂C₂@C₂(61)-C₉₂. Importantly, this is the first crystallographic identification of di-yttrium EMFs, Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂, featuring unsupported Y–Y bonds. Prior to our work, direct yttrium–yttrium contacts have only been observed in the interstitial compounds Y₄I₅C and Y₆I₇C₂ (ref. 28) but never in any organometallic complexes. Interestingly, our results show that the cluster configuration changes from Y₂ to Y₂C₂ as the cage enlarges accompanied by the donation of the valence electrons forming a Y–Y bond with the inserted C₂-unit, and Y₂C₂@C_2n (2n = 86–92) are finally formed.

Results and discussion

Soot containing Y₂C_2n (2n = 82, 88–94) isomers was synthesized by a direct-current arc discharge method and Y₂C_2n isomers were isolated via a combinational process involving Lewis-acid treatment and high performance liquid chromatography (HPLC) separation (Fig. S1–S5†). Experimental details are provided in the ESI.† The analytical HPLC chromatograms (Fig. S6†) and the laser-desorption/ionization time-of-flight (LDI-TOF) mass spectra (Fig. S7†) of Y₂C_2n (2n = 82, 88–94) reveal their high purity which guarantee the following characterization.

Co-crystals of Y₂C_2n (2n = 82, 88–94) with Ni^II(OEP) (OEP = 2,3,7,8,12,13,17,18-octaethylporphyrin dianion) were obtained by layering a benzene solution of Ni^II(OEP) over the CS₂ solution of each Y₂C_2n isomer, and were used to unambiguously determine their molecular structures by single-crystal X-ray diffraction (XRD) crystallography. Details of the crystallographic data are listed in Table S1.† The results unambiguously confirm that two of them are di-EMFs with relatively small cages, namely Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂, whereas the others are all carbide cluster metallofullerenes (CCMFs), i.e., Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂, respectively. Although the C_s(6)-C₈₂, C_3v(8)-C₈₂ and C₂(41)-C₉₀ cages have been observed for other EMFs, such as Lu₂@C_s(6)-C₈₂, Lu₂@C_3v(8)-C₈₂ and La₂C₂@C₂(41)-C₉₀,^26,29 the C_s(15)-C₈₆, C₁(26)-C₈₈ and C₂(61)-C₉₂ cages are unprecedented.

Inside the fullerene cages, the yttrium atoms show some degree of disorder. Details of the disorder are illustrated in Fig. 1 and the occupancy values are summarized in Table S2.† In the two di-EMFs, 8 and 14 yttrium sites were found for the two Y atoms in C_s(6)-C₈₂ and C_3v(8)-C₈₂, respectively, indicating a motional behavior of the Y atom to pursue strong yttrium–cage interactions. As for the CCMFs, i.e., Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂, the number of the disordered yttrium sites increases along with cage expansion (Fig. 1). In detail, the Y₂C₂ unit displays 4, 4, 8 and 17 sites in C_s(15)-C_86,C₁(26)-C₈₈, C₂(41)-C₉₀ and C₂(61)-C₉₂, respectively. It appears that the motional behavior of the metal atoms inside fullerene cages is an effective way to ensure sufficient metal–cage/metal–C₂ unit interactions.

Fig. 1 Perspective drawings showing the internal yttrium sites. (a) 8 in Y₂@C_s(6)-C₈₂, (b) 14 in Y₂@C_3v(8)-C₈₂, (c) 4 in Y₂C₂@C_s(15)-C₈₆, (d) 4 in Y₂C₂@C₁(26)-C₈₈, (e) 8 in Y₂C₂@C₂(41)-C₉₀, and (f) 17 in Y₂C₂@C₂(61)-C₉₂.

Moreover, the representative structural data of Y₂C_2n (2n = 82, 88–94) isomers, such as the structural parameters of the internal species, and Ni–cage and Y–cage distances, are summarized in Table S3.† In detail, the shortest Y–cage distances are 2.321 Å, 2.357 Å, 2.203 Å, 2.041 Å, 2.010 Å and 2.042 Å for Y₂@C_s(6)-C₈₂, Y₂@C_3v(8)-C₈₂, Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀ and Y₂C₂@C₂(61)-C₉₂, respectively. These values are comparable to those observed for the derivatives of Y₂C₂@C_s(6)-C₈₂,³⁰ suggesting strong Y–cage interactions. Moreover, the distances between the Y ions and the internal C₂ units for these CCMFs range from 2.034 Å to 2.711 Å (Table S3†), which are similar to that observed for Y₂C₂@C₁(1660)-C₁₀₈, representing typical coordination bonds.³¹

Fig. 2 presents the molecular structures of these EMFs showing the major components together with the co-crystallized Ni^II(OEP) molecule. The shortest Ni–cage distances range from 2.725 Å to 3.027 Å, suggesting substantial π–π interactions between the fullerene cage and Ni^II(OEP). For M₂@C_2n-type EMFs, Popov et al. have suggested theoretically that the internal metals may not adopt their highest oxidation states, thus yielding a possibility of covalent metal–metal bonding in such EMFs as Lu₂@C₇₆ and M₂@C₈₂ (M = Sc, Y, Er, Lu, etc.).^20,25,32 Indeed, our concrete crystallographic results reveal that Lu₂@C_2n (2n = 82–86) are all di-EMFs with a Lu–Lu single bond formed between the two internal lutetium ions.²⁶ Consistently, formation of the Y–Y bond is also confirmed by our crystallographic results of the two di-EMFs. The Y–Y distances between the major Y sites are 3.635 Å and 3.596 Å for Y₂@C_3v(6)-C₈₂ and Y₂@C_s(8)-C₈₂, respectively, which are comparable to the calculated Y–Y single bond length (3.695 Å) for Y₂@C₈₂ isomers,^20,32 confirming that each Y atom adopts a low divalent state of +2 and the third valence electron is donated to form the metal–metal bond.

Fig. 2 ORTEP drawings of (a) Y₂@C_s(6)-C₈₂, (b) Y₂@C_3v(8)-C₈₂, (c) Y₂C₂@C_s(15)-C₈₆, (d) Y₂C₂@C₁(26)-C₈₈, (e) Y₂C₂@C₂(41)-C₉₀, and (f) Y₂C₂@C₂(61)-C₉₂. Only one cage orientation and the predominant cluster site are shown, while minor sites and solvent molecules are omitted for clarity.

As for the CCMFs, the Y–Y distances between the two major Y sites are 4.121 Å, 4.271 Å, 4.024 Å, and 4.349 Å, for Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂, respectively (Table S3†), which are all much longer than the theoretical value of an Y–Y single bond (3.695 Å),^20,32 thus enabling the insertion of a C₂-unit to coordinate with the two Y atoms. In addition, the C–C distances of the C₂ unit in Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂ are 1.075 Å, 1.131 Å, 1.052 Å, and 1.046 Å, respectively, indicative of typical C–C triple bonds (Fig. S8 and Table S3†).^33–36

Fig. 3 shows the visible-near-infrared (Vis-NIR) absorption spectra of the EMFs dissolved in carbon disulfide (CS₂) under study, all showing characteristic bands in the range of 400–1400 nm, as summarized in Table S4.† Specifically, the absorption onsets result in a large optical bandgap for Y₂@C_3v(8)-C₈₂ (1.13 eV), small bandgaps for Y₂C₂@C_s(15)-C₈₆ (0.72 eV) and Y₂C₂@C₂(41)-C₉₀ (0.86 eV), and moderate bandgaps for Y₂@C_s(6)-C₈₂ (0.90 eV), Y₂C₂@C₁(26)-C₈₈ (0.91 eV) and Y₂C₂@C₂(61)-C₉₂ (0.98 eV).

Fig. 3 Vis-NIR absorption spectra of Y₂C_2n (2n = 82, 88–94) isomers dissolved in CS₂.

The two di-EMFs, namely Y₂@C_3v(8)-C₈₂ and Y₂@C_s(6)-C₈₂, and the largest CCMF, Y₂C₂@C₂(61)-C₉₂, are chosen as representatives for the study of their electrochemical properties by cyclic voltammetry (Fig. S9†), whereas the redox behaviors of the other EMFs under study have not been obtained due to their limited amounts. In general, these three EMFs exhibit one reversible oxidation step and four reversible reduction processes (Fig. S9†). The detailed redox potentials are given in Table 1 along with the values of related EMFs for comparison. It is noteworthy that the first oxidation potentials for the two di-EMFs, i.e., Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂, are identical (−0.16 V), which are much lower than that of Y₂C₂@C₂(61)-C₉₂ (0.28 V) as well as those of the other di-EMFs reported previously,^22,26,37,38 such as Sc₂@C_3v(8)-C₈₂ (0.05 V), Lu₂@C_s(6)-C₈₂ (0.34 V) and La₂@I_h(7)-C₈₀ (0.56 V), revealing their high electron affinity. These results are in good agreement with the speculation by Popov et al. and show the metal dependence of the first oxidation potential for di-EMFs with the same cages, which is rationalized through analysis of the energy of M–M bonding molecular orbitals because the electron should be removed from M–M bonding orbitals in the first oxidation process.³² Hence the lowest first oxidation potential of Y₂@C₈₂ (e.g., C_3v(8)- or C_s(6)-C₈₂) among M₂@C₈₂ (M = Sc, Y, Lu and Er) is probably ascribed to the high energy of the Y–Y bonding orbital.²⁰ In addition, the redox potentials of Y₂@C_3v(8)-C₈₂ are cathodically shifted as compared to the corresponding values of M₂@C_3v(8)-C₈₂ (M = Sc, Er and Lu),³² indicating that the electrochemical properties of EMFs can be readily manipulated by adjusting the encapsulated species inside the cages.

Table 1 Redox potentials (V vs. Fc/Fc⁺)^a of Y₂@C_3v(8)-C₈₂, Y₂@C_s(6)-C₈₂ and Y₂C₂@C₂(61)-C₉₂ along with those of related EMFs

EMFs	^ox E 1	^red E 1	^red E 2	^red E 3	^red E 3	E gap	Ref.
a Half-cell potentials are given unless otherwise stated. b E gap = (^oxE1 − ^redE1).
Y2@C3v(8)-C82	−0.16	−1.23	−1.61	−2.18	−2.46	1.07	This work
Sc2@C3v(8)-C82	0.02	−1.16	−1.53	−1.73	−2.02	1.18	32
Er2@C3v(8)-C82	0.13	−1.14	−1.41	−1.83	—	1.27	32
Lu2@C3v(8)-C82	0.50	−1.16	−1.46	−1.77	—	1.66	32
Y2@Cs(6)-C82	−0.16	−1.06	−1.39	−1.85	−2.15	0.90	This work
Er2@Cs(6)-C82	0.02	−1.01	−1.31	—	—	1.03	32
Lu2@Cs(6)-C82	0.34	−1.02	−1.35	−1.77	—	1.36	26
Y2C2@C2(61)-C92	0.28	−1.00	−1.49	−1.73	−1.98	1.28	This work

---

Density functional theory (DFT) calculations at the M06-2X/6-31G*∼SDD level were conducted to rationalize the formation of these stable Y₂C_2n (2n = 82, 88–94) isomers. Fig. S10† depicts their optimized geometries, which agree well with the X-ray structures. For the two di-EMFs, the optimized Y–Y distances are 3.60 Å and 3.54 Å for Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂, respectively, which are consistent with our crystallographic results, clearly indicating a direct Y–Y bond between the two yttrium ions. Moreover, it was proposed that both the C_s(6)-C₈₂ and C_3v(8)-C₈₂ cages have a low-lying LUMO and LUMO+1, but a high-lying LUMO+2, whose energies are higher than that of the Y–Y bonding MO.²⁰ Therefore, the two Y atoms ([Kr]4d¹5s²) tend to adopt the divalent state with the remaining two electrons paring to generate an Y–Y single bond. Indeed, our calculations for Y₂, C_s(6)-C₈₂, C_3v(8)-C₈₂, Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂ confirm that there are large energy gaps between the LUMO+2 of the two hollow cages and the HOMOs of the Y₂ dimer, implying unfavorable electron transfer from the latter to the former (Fig. S11†). Further natural bond orbital (NBO) analysis demonstrates that the two Y atoms in the two Y₂@C₈₂ isomers form an Y–Y single bond with an electron occupancy of 1.97 e, which is supported by the calculated Wiberg bond orders (WBOs) for Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂ ranging from 1.11 to 1.12 (Table S5†). The Y–Y bonds have spd-hybrid character with the Y-5s orbitals contributing the most to the metal bonding MOs, and each Y atom donates one 4d electron and one 5s electron to the cage.

As for the CCMFs, however, the calculated Y–Y distances in Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂ are as long as 4.53 Å, 4.25 Å, 4.40 Å and 4.56 Å, respectively, which are consistent with the X-ray diffraction values, indicating the absence of direct Y–Y bonding. Consistently, the calculated WBO values for these CCMFs range from 0.29 to 0.41 (Table S5†), revealing clearly that the valence electrons forming the Y–Y bond in di-EMFs are donated to the C₂-unit to form the CCMFs.

Further computational studies were done on different low-lying Y₂C_2n (2n = 82, 88–94) isomers in either the Y₂@C_2n or Y₂C₂@C_2n−2 form to rationalize the existence of the experimentally obtained Y₂C_2n isomers. Fig. S12† to S16 show their optimized structures and relative energies as well as HOMO–LUMO gap energies. Since fullerenes and EMFs are formed at very high temperatures (1500–3000 K) under arc discharge conditions,³⁹ we have also analyzed the relative stability of the most favorable EMFs in terms of their relative Gibbs free energies.⁴⁰ As clearly shown in Fig. S12,†Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂ are the lowest-energy ones among all the considered isomers, and Fig. 4a clearly shows that they are the most abundant isomers in the temperature range for EMF formation. In comparison, when the number of carbon atoms increases from C₈₈ to C₉₀, our calculations reveal that the experimentally obtained CCMF isomers, Y₂C₂@C_s(15)-C₈₆ and Y₂C₂@C₁(26)-C₈₈, are more than 8.2 kcal mol⁻¹ higher in energy than the corresponding lowest-lying di-EMFs isomers (Fig. S13 and S14†). However, when the Gibbs energy is taken into account, they become the most abundant isomers above 1500 K, with the molar fractions of the corresponding di-EMFs being almost negligible (Fig. 4b and c).

Fig. 4 Computed molar fraction as a function of temperature for the low-lying isomers of (a) Y₂C₈₂, (b) Y₂C₈₈, (c) Y₂C₉₀, (d) Y₂C₉₂, and (e) Y₂C₉₄. The isomers marked with red boxes are experimentally obtained.

As for Y₂C₉₂₋₉₄, Y₂C₂@C₂(41)-C₉₀, Y₂C₂@C₂(61)-C₉₂ and their corresponding di-EMF isomers are mixed in terms of their potential energy, and the most stable Y₂C₉₂ and Y₂C₉₄ are both CCMFs, namely, Y₂C₂@C₂(41)-C₉₀ and Y₂C₂@D₃(85)-C₉₂, respectively (Fig. S15 and S16†). From Fig. 4d, we can see that Y₂C₂@C₂(41)-C₉₀ is the dominant species in the whole range of temperatures up to 5000 K. Moreover, as for Y₂C₉₄, although Y₂C₂@D₃(85)-C₉₂ is the dominant species at low temperatures, the concentration of the experimentally obtained Y₂C₂@C₂(61)-C₉₂ rapidly increases when the temperature rises, finally surpassing that of Y₂C₂@D₃(85)-C₉₂ after 3400 K (Fig. 4e).

Overall, our experimental and theoretical results have unambiguously confirmed that the Y₂ dimer chose to form a Y–Y single bond in relatively small cages like C₈₂, but the cluster configuration changes from Y₂ to Y₂C₂ when the cage expands, thus forming the CCMFs in large cages C_86–92. It is inferred that larger metals such as erbium and lanthanum may prefer a larger cage to form the corresponding CCMFs. For example, Stevenson et al. reported the isolation and crystallographic characterization of an erbium-based CCMF with a giant C₉₂ cage, namely Er₂C₂@D₃(85)-C₉₂.⁴¹ Moreover, our group found that the even larger La³⁺ ions tend to form CCMFs with some giant cages such as C_90–104.^29,35,36 In addition, the cluster configuration is another critical factor that may affect the cage size. For instance, a linear Y₂C₂ cluster can be encapsulated into the giant C₁₀₈ cage to template the formation of Y₂C₂@C₁(1660)-C₁₀₈.³¹ There is still a long way to finally clarify the mysteries in the metal–metal interactions in EMFs and new theories of coordination chemistry are expected.

Conclusions

A series of di-yttrium EMFs Y₂C_2n (2n = 82, 88–94) have been successfully isolated and fully characterized by mass spectrometry, UV-vis-NIR and single-crystal X-ray crystallography. The crystallographic results unambiguously confirm that two of them are di-EMFs, namely Y₂@C_s(6)-C₈₂ and Y₂@C_3v(8)-C₈₂, whereas the others are all CCMFs, i.e., Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₁(26)-C₈₈, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂, respectively. It is noteworthy that our experimental and theoretical studies clearly reveal that the Y₂ dimer preferentially forms a Y–Y single bond in the di-EMFs with small cages, and further cage expansion alters its configuration to Y₂C₂ because the valence electrons forming the Y–Y bond are donated to the inserted C₂-unit, thus resulting in the formation of the corresponding CCMFs in larger cages. In addition, our computational results demonstrate that the high thermodynamic stability of all these EMFs originates from a four-electron transfer and the bonding nature between the internal metallic species and cages. Accordingly, this work has confirmed for the first time the existence of the Y–Y bond inside fullerene cages and the transformation from di-EMFs to CCMFs along with the cage expansion, presenting an in-depth understanding and new sights into the metal–metal/metal–carbon interactions in such rarely explored metal–carbon hybrid molecules.

Experimental section

Synthesis of Y₂C_2n (2n = 82, 88–94) isomers

Soot containing yttrium-EMFs was synthesized using a direct-current arc discharge method. Briefly, a core-drilled graphite rod filled with graphite/Y₂O₃ (molar ratio: Y/C = 1 : 50) was burned under a 225 Torr helium atmosphere with a power of 100 A × 20 V. The soot was then extracted with CS₂. After removal of CS₂, the residue was dissolved in toluene and the solution was subjected to high performance liquid chromatography (HPLC) for the subsequent separation. After a combinational process involving Lewis-acid treatment and multi-stage HPLC separation, six pure isomers of yttrium-EMFs are finally obtained. Experimental details are described in the ESI.†

Co-crystallization and crystallographic solutions of Y₂C_2n (2n = 82, 88–94) cocrystals

Crystalline blocks of Y₂C₂@C_2n·Ni(OEP) were obtained by layering a benzene solution of Ni(OEP) over a CS₂ solution of the corresponding metallofullerene at room temperature for two weeks. Single-crystal XRD measurements of Y₂C₂@C₁(26)-C₈₈ and Y₂@C_3v(8)-C₈₂ were performed at 173 K on a Bruker D8 QUEST machine equipped with a CMOS camera (Bruker AXS Inc., Germany), and X-ray data of Y₂@C_s(6)-C₈₂ were obtained at 100 K at the BL17U station of the Shanghai Synchrotron Radiation Facility. Crystallographic characterization of the other three EMFs, i.e., Y₂C₂@C_s(15)-C₈₆, Y₂C₂@C₂(41)-C₉₀, and Y₂C₂@C₂(61)-C₉₂, was performed at 100 K at the BL17B station of the Shanghai Synchrotron Radiation Facility. The multi-scan method was used for absorption corrections. The structures were solved by the direct method and were refined with SHELXL-2014/7.⁴² Co-crystals of Y₂@C_3v(8)-C₈₂·Ni^II(OEP) and Y₂C₂@C₂(61)-C₉₂·Ni^II(OEP) contain a severely disordered lattice of C₆H₆ and CS₂ molecules that could not be modeled properly. Therefore, the SQUEEZE program, a part of the PLATON package of crystallographic software,^43,44 was used to calculate the solvent disorder area and remove its contribution from the intensity data. CCDC-1811906 (Y₂@C_s(6)-C₈₂), CCDC-1812333 (Y₂@C_3v(8)-C₈₂), CCDC-1528591 (Y₂C₂@C_s(15)-C₈₆), CCDC-1528590 (Y₂C₂@C₁(26)-C₈₈), CCDC-1528691 (Y₂C₂@C₂(41)-C₉₀), and CCDC-1812135 (Y₂C₂@C₂(61)-C₉₂) contain the supplementary crystallographic data for this paper.†

Computational details

Density functional theory calculations were carried out by using the M06-2X⁴⁵ functional in conjunction with the 6-31G* basis set for C^46,47 and SDD basis set and the corresponding effective core potential for Y⁴⁸ (denoted as 6-31G*∼SDD), as implemented in the Gaussian 09 software package.⁴⁹ More details are given in the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Financial support from the NSFC (No. 51672093, 51602112, 51602097 and 21103224) is gratefully acknowledged. We thank the staff of the BL17B & BL17U beamline of National Facility for Protein Science Shanghai (NFPS) at the Shanghai Synchrotron Radiation Facility for assistance during data collection. Also, we are grateful to the Analytical and Testing Center in Huazhong University of Science and Technology for all related measurements.

Notes and references

1. M. S. Hill, P. B. Hitchcock and R. Pongtavornpinyo, Science, 2006, 311, 1904–1907.
2. M. V. Butovskii and R. Kempe, New J. Chem., 2015, 39, 7544–7558.
3. S. T. Liddle and D. P. Mills, Dalton Trans., 2009, 5592–5605.
4. L. Bao, P. Peng and X. Lu, Acc. Chem. Res., 2018, 51, 810–815.
5. A. A. Popov, S. Yang and L. Dunsch, Chem. Rev., 2013, 113, 5989–6113.
6. X. Lu, L. Feng, T. Akasaka and S. Nagase, Chem. Soc. Rev., 2012, 41, 7723.
7. X. Zhang, Y. Wang, R. Morales-Martínez, J. Zhong, C. de Graaf, A. Rodríguez-Fortea, J. M. Poblet, L. Echegoyen, L. Feng and N. Chen, J. Am. Chem. Soc., 2018, 140, 3907–3915.
8. W. Cai, R. Morales-Martínez, X. Zhang, D. Najera, E. L. Romero, A. Metta-Magaña, A. Rodríguez-Fortea, S. Fortier, N. Chen, J. M. Poblet and L. Echegoyen, Chem. Sci., 2017, 8, 5282–5290.
9. T.-S. Wang, N. Chen, J.-F. Xiang, B. Li, J.-Y. Wu, W. Xu, L. Jiang, K. Tan, C.-Y. Shu, X. Lu and C.-R. Wang, J. Am. Chem. Soc., 2009, 131, 16646–16647.
10. H. Fang, H. Cong, M. Suzuki, L. Bao, B. Yu, Y. Xie, N. Mizorogi, M. M. Olmstead, A. L. Balch, S. Nagase, T. Akasaka and X. Lu, J. Am. Chem. Soc., 2014, 136, 10534–10540.
11. C.-H. Chen, K. B. Ghiassi, M. R. Cerón, M. A. Guerrero-Ayala, L. Echegoyen, M. M. Olmstead and A. L. Balch, J. Am. Chem. Soc., 2015, 137, 10116–10119.
12. T. Cai, L. Xu, M. R. Anderson, Z. Ge, T. Zuo, X. Wang, M. M. Olmstead, A. L. Balch, H. W. Gibson and H. C. Dorn, J. Am. Chem. Soc., 2006, 128, 8581–8589.
13. S. Stevenson, H. M. Lee, M. M. Olmstead, C. Kozikowski, P. Stevenson and A. L. Balch, Chem.–Eur. J., 2002, 8, 4528–4535.
14. N. Chen, M. N. Chaur, C. Moore, J. R. Pinzón, R. Valencia, A. Rodríguez-Fortea, J. M. Poblet and L. Echegoyen, Chem. Commun., 2010, 46, 4818–4820.
15. S. Stevenson, M. A. Mackey, M. A. Stuart, J. P. Phillips, M. L. Easterling, C. J. Chancellor, M. M. Olmstead and A. L. Balch, J. Am. Chem. Soc., 2008, 130, 11844–11845.
16. B. Q. Mercado, M. M. Olmstead, C. M. Beavers, M. L. Easterling, S. Stevenson, M. A. Mackey, C. E. Coumbe, J. D. Phillips, J. Paige Phillips, J. M. Poblet and A. L. Balch, Chem. Commun., 2010, 46, 279–281.
17. Q. Tang, L. Abella, Y. Hao, X. Li, Y. Wan, A. Rodríguez-Fortea, J. M. Poblet, L. Feng and N. Chen, Inorg. Chem., 2015, 54, 9845–9852.
18. T.-S. Wang, L. Feng, J.-Y. Wu, W. Xu, J.-F. Xiang, K. Tan, Y.-H. Ma, J.-P. Zheng, L. Jiang, X. Lu, C.-Y. Shu and C.-R. Wang, J. Am. Chem. Soc., 2010, 132, 16362–16364.
19. F. Liu, C.-L. Gao, Q. Deng, X. Zhu, A. Kostanyan, R. Westerström, S. Wang, Y.-Z. Tan, J. Tao, S.-Y. Xie, A. A. Popov, T. Greber and S. Yang, J. Am. Chem. Soc., 2016, 138, 14764–14771.
20. A. A. Popov, S. M. Avdoshenko, A. M. Pendás and L. Dunsch, Chem. Commun., 2012, 48, 8031.
21. T. Zuo, L. Xu, C. M. Beavers, M. M. Olmstead, W. Fu, T. D. Crawford, A. L. Balch and H. C. Dorn, J. Am. Chem. Soc., 2008, 130, 12992–12997.
22. L. Bao, M. Chen, C. Pan, T. Yamaguchi, T. Kato, M. M. Olmstead, A. L. Balch, T. Akasaka and X. Lu, Angew. Chem., Int. Ed., 2016, 55, 4242–4246.
23. M. Yamada, H. Kurihara, M. Suzuki, M. Saito, Z. Slanina, F. Uhlik, T. Aizawa, T. Kato, M. M. Olmstead, A. L. Balch, Y. Maeda, S. Nagase, X. Lu and T. Akasaka, J. Am. Chem. Soc., 2015, 137, 232–238.
24. F. Liu, D. S. Krylov, L. Spree, S. M. Avdoshenko, N. A. Samoylova, M. Rosenkranz, A. Kostanyan, T. Greber, A. U. B. Wolter, B. Büchner and A. A. Popov, Nat. Commun., 2017, 8, 16098.
25. T. Yang, X. Zhao and E. Osawa, Chem.–Eur. J., 2011, 17, 10230–10234.
26. W. Shen, L. Bao, Y. Wu, C. Pan, S. Zhao, H. Fang, Y. Xie, P. Jin, P. Peng, F.-F. Li and X. Lu, J. Am. Chem. Soc., 2017, 139, 9979–9984.
27. W. Shen, L. Bao, S. Hu, L. Yang, P. Jin, Y. Xie, T. Akasaka and X. Lu, Chem. Sci., 2019, 10, 829–836.
28. S. M. Kauzlarich, T. Hughbanks, J. D. Corbett, P. Klavins and R. N. Shelton, Inorg. Chem., 1988, 27, 1791–1797.
29. S. Zhao, P. Zhao, W. Cai, L. Bao, M. Chen, Y. Xie, X. Zhao and X. Lu, J. Am. Chem. Soc., 2017, 139, 4724–4728.
30. F. Jin, N. B. Tamm, S. I. Troyanov and S. Yang, J. Am. Chem. Soc., 2018, 140, 3496–3499.
31. C. Pan, L. Bao, X. Yu, H. Fang, Y. Xie, T. Akasaka and X. Lu, ACS Nano, 2018, 12, 2065–2069.
32. N. A. Samoylova, S. M. Avdoshenko, D. S. Krylov, H. R. Thompson, A. C. Kirkhorn, M. Rosenkranz, S. Schiemenz, F. Ziegs, A. U. B. Wolter, S. Yang, S. Stevenson and A. A. Popov, Nanoscale, 2017, 23, 7977–7990.
33. C.-H. Chen, L. Abella, M. R. Cerón, M. A. Guerrero-Ayala, A. Rodríguez-Fortea, M. M. Olmstead, X. B. Powers, A. L. Balch, J. M. Poblet and L. Echegoyen, J. Am. Chem. Soc., 2016, 138, 13030–13037.
34. H. Yang, C. Lu, Z. Liu, H. Jin, Y. Che, M. M. Olmstead and A. L. Balch, J. Am. Chem. Soc., 2008, 130, 17296–17300.
35. W. Cai, F.-F. Li, L. Bao, Y. Xie and X. Lu, J. Am. Chem. Soc., 2016, 138, 6670–6675.
36. W. Cai, L. Bao, S. Zhao, Y. Xie, T. Akasaka and X. Lu, J. Am. Chem. Soc., 2015, 137, 10292–10296.
37. M. Yamada, M. Minowa, S. Sato, Z. Slanina, T. Tsuchiya, Y. Maeda, S. Nagase and T. Akasaka, J. Am. Chem. Soc., 2011, 133, 3796–3799.
38. H. Kurihara, X. Lu, Y. Iiduka, N. Mizorogi, Z. Slanina, T. Tsuchiya, S. Nagase and T. Akasaka, Chem. Commun., 2012, 48, 1290–1292.
39. W. Krätschmer, L. D. Lamb, K. Fostiropoulos and D. R. Huffman, Nature, 1990, 347, 354–358.
40. Z. Slanina, S.-L. Lee, F. Uhlík, L. Adamowicz and S. Nagase, Theor. Chem. Acc., 2007, 117, 315–322.
41. A. L. Balch, M. M. Olmstead, X. Powers, C. M. Davison, S. Stevenson and K. R. Tepper, Chem.–Eur. J., 2018, 24, 13479–13484.
42. G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008, 64, 112–122.
43. A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13.
44. P. van der Sluis and A. L. Spek, Acta Crystallogr., Sect. A: Found. Crystallogr., 1990, 46, 194–201.
45. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
46. G. A. Petersson and M. A. Al-Laham, J. Chem. Phys., 1991, 94, 6081–6090.
47. G. A. Petersson, A. Bennett, T. G. Tensfeldt, M. A. Al-Laham, W. A. Shirley and J. Mantzaris, J. Chem. Phys., 1988, 89, 2193–2218.
48. X. Cao and M. Dolg, J. Mol. Struct.: THEOCHEM, 2002, 581, 139–147.
49. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci and G. A. Petersson, et al., Gaussian 09, Gaussian, Inc., Wallingford, CT, USA, 2009.

---

Footnotes

† Electronic supplementary information (ESI) available. CCDC 1528590, 1528591, 1528691, 1811906, 1812135 and 1812333. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9sc00941h

‡ These authors contributed equally to this work.

---