Electronic Properties and ¹³C NMR Structural Study of Y₃N@C₈₈

Abstract

In this paper we report the synthesis, purification, ¹³C NMR and other characterization studies of Y₃N@C₈₈. The ¹³C NMR, UV-vis and chromatographic data suggest an Y₃N@C₈₈ having IPR allowed cage with D₂(35)-C₈₈ symmetry. In earlier density functional theory (DFT) computational and X-ray crystallographic studies, it was reported that lanthanide (A₃N)⁶⁺ clusters are stabilized in D₂(35)-C₈₈ symmetry cages and have reduced HOMO-LUMO gaps relative to other trimetallic nitride endohedral metallofullerene cage systems, for example, A₃N@C₈₀. In this paper, we report that the non-lanthanide (Y₃N)⁶⁺ cluster in the D₂(35)-C₈₈ cage exhibits a HOMO-LUMO gap consistent with other lanthanide A₃N@C₈₈ molecules based on electrochemical measurements and DFT computational study. These results suggests that the reduced HOMO-LUMO gap of A₃N@C₈₈ systems is a property dominated by the D₂(35)-C₈₈ carbon cage and not f-orbital lanthanide electronic metal cluster (A₃N)⁶⁺ orbital participation.

Introduction

The separation and characterization of higher fullerenes beyond C₈₀ is very difficult, not only because of their extremely low yield, but also the possible presence of multiple isomers, instability, and lower solubility.¹ For example, Yang et al. reported the separation of Di-and Tridysprosium Endohedral Metallofullerenes in the fullerene cages from C₉₄ to C₁₀₀.² Due to the limited available amount of sample, they did not obtain any structural information of those large cages. ² Among the large fullerene cages, the C₈₈ cage has been widely studied. As early as 1995, C₈₈ empty cage fullerenes were separated from soot generated in a Kratschmer-Huffman electric-arc generator and subsequently characterized using ¹³C NMR by Achiba et al.³ In a later report, Miyake et al.⁴ described a more detailed ¹³C NMR study for the C₈₈ empty cage fullerene. However, the structural assignments were ambiguous because at least three C₈₈ isomers exist.⁴ Due to the difficulty in experimental part, the extensive theoretical studies were performed by several groups to aid the structural determination.^5,6 It was predicted that C₂(7), C_s(17) and C₂(33) isomers are the mostly likely to be experimentally-isolated isomers among the 35 IPR-obeying isomers of C₈₈.^5,6 Derivatization of empty fullerene cages has proved to be an effective tool for cage stabilization and meanwhile provides a path for further structural characterization. For example, Troyanov and Tamm ⁷ synthesized and characterized trifluoromethyl derivatives of C₈₈, C₈₈(CF₃)₁₈. Based on their X-ray investigation, they determined that C₈₈(CF₃)₁₈ processes a C₂(33)-C₈₈ cage, which is one of the three most stable isomers predicted on a theoretical basis. Encapsulation of atoms or clusters into fullerene cages is another method used to stabilize empty fullerene cages by electron transfer processes between the encapsulated and the carbon cage.

Trimetallic nitride template (TNT) endohedral metallofullerenes (EMFs) are particularly important not only because of their relatively high yields, but also because of their intriguing potential applications in biomedical, optoelectronic, and photovoltaic fields.^8,9 Insertion of a TNT cluster into a C₈₈ cage (M₃N@C₈₈, M= Gd, Tm, Dy, Tb, Nd, Pr, Ce, La) has been reported by several laboratories.^10–13 Echegoyen and coworkers suggested that the C₈₈ cage is preferentially templated by a TNT cluster for metal ions with ionic radius larger than gadolinium, such as, Nd, Pr, Ce.^10,11 However, most of the TNT EMFs with the C₈₈ cage have been characterized by mass spectroscopy without definitive structural characterization. Notable exceptions are the single crystal X-ray studies reported for Tb₃N@C₈₈ which have been found to have an IPR-obeying D₂(35) structure.¹² This is in contrast with the theoretical predictions for the empty C₈₈ cage. Thus, it is interesting to examine whether clusters with different metals (A₃N)⁶⁺ encapsulated within a C₈₈ cage have the same cage structure as IPR-obeying Tb₃N@D₂(35)-C₈₈.

In this paper, we report the synthesis, separation and structural characterization of diamagnetic Y₃N@C₈₈. This diamagnetic molecule allows high-resolution ¹³C NMR structural studies, which are generally not feasible for the paramagnetic lanthanide C₈₈ cage TNT EMFs reported above. Density function theory calculations were utilized to augment the structural determination. The electronic properties of Y₃N@C₈₈ were also studied by both electrochemical and DFT computational approaches to compare a non-lanthanide TNT EMF with the corresponding lanthanide TNT EMFs reported to date.

Experimental Section

A sample of Y₃N@C₈₈ was synthesized in an electric arc-discharge reactor by vaporizing graphite rods containing a mixture of Y₂O₃ and graphite powder and using Cu as catalyst with a weight ratio of 1.1:1.0:2.1 in a dynamic flow of N₂ and He (flow rate ratio of N₂/He=3:100).¹⁴ The toluene extract from the raw soot was applied to a cyclopentadiene-functionalized Merrifield peptide resin. The eluent was further separated by two-stage HPLC. The first stage was carried out on a 5PBB column and the sixth fraction contains Y₃N@C₈₈, which was further purified by a 5PYE column.

The ¹³C NMR measurements (150 MHz) were performed on a Bruker Advance spectrometer (600 MHz, ¹H). The sample was dissolved in CS₂ with Cr(acac)₃ as the relaxation agent and acetone-d₆ as the internal lock at 25 °C. Cyclic voltammetry was conducted using a CH Instruments 600A potentiostat (Austin, TX) with a single-compartment, three electrode, electrochemical cell. A 2.0 mm glassy carbon working electrode, platinum wire auxiliary, and a silver wire pseudo-reference electrode was used; ferrocene was used as an internal standard.

Density functional theory (DFT) computations were performed using the Gaussian 03 program package. All of the molecules were geometrically optimized at the UB3LYP level with a DZVP basis set for yttrium atoms and a 6-31G* basis set for carbon and nitrogen atoms.¹⁴ DFT-optimized energy values were obtained starting from the X-ray crystallographic structures of the corresponding Tb₃N@D₂(35)-C₈₈.¹¹

Results and Discussion

As illustrated in Figure 1, the HPLC, positive ion LD-TOF MS, and UV-vis spectra for Y₃N@C₈₈ is compared with data that was previously reported for Tb₃N@C₈₈.¹² The close correspondence of the data in Figure 1 strongly suggests that Y₃N@C₈₈ has the same D₂(35)cage structure as Tb₃N@D₂(35)-C₈₈.¹² Because the Y₃N@C₈₈ molecule is diamagnetic, we were able to obtain high-resolution ¹³C NMR data for Y₃N@C₈₈ to further confirm the cage structure.

Figure 1.

(a) HPLC chromatogram of the purified Y₃N@C₈₈ and Tb₃N@C₈₈ (10 × 250 mm 5PYE column; λ=390 nm; flow rate 2.0 mL/min; toluene as eluent; 25 °C. (b) UV-vis spectra of Y₃N@C₈₈ and Tb₃N@C₈₈ in toluene and positive ion LD-TOF MS for purified Y₃N@C₈₈.

The ¹³C NMR spectrum (Figure 2a) for Y₃N@C₈₈ exhibits a total of 22 lines (lines at 139.10 and 139.69 ppm, double intensity) with a shift range from 131.0 to 150.0 ppm. This result is completely consistent with the 22 ¹³C NMR signals observed for Y₃N@C₈₈. An IPR-allowed isomer is suggested because of the absence of a ¹³C NMR signal above 155 ppm, which is characteristic of non-IPR pentalene motifs.¹⁴ There are 35 IPR-allowed C₈₈ isomers and two isomers (1) and (35) that have D₂ symmetry (22 × 4 lines) which is consistent with the observed spectrum.¹⁵ Among the 35 IPR-obeying isomers for the C₈₈ cage, no other IPR-allowed structure would exhibit fewer than 22 lines except for isomer (34) with T symmetry that would exhibit 8 ¹³C NMR resonances (1×4 and 7×12). All other C₈₈ IPR isomers have lower symmetry and would exhibit more than 22 spectral lines. In addition, there are 7 pyrene 6,6,6 type carbons that range from 132.0–137.9 ppm, which is in reasonably good agreement with 7 DFT predicted values (Figure 2b) ranging from 131.9–139.2 ppm.¹⁴ The correlation between experimental ¹³C and the DFT predicted chemical shielding values are shown in Figure 2c. Tb₃N@D₂(35)-C₈₈ exhibits a carbon cage with D₂ symmetry as determined by previous single-crystal X-ray diffraction studies. Thus, our current results are consistent with a Y₃N@D₂(35)-C₈₈ structure in analogous fashion to Tb₃N@D₂(35)-C₈₈.¹²

Figure 2.

(a)The ¹³C NMR spectrum of Y₃N@D₂-C₈₈ in CS₂ with 10 mg Cr(acac)₃ relaxant, acetone-d₆ lock) after 64,000 scan at 25 °C, showing the 22×4 pattern (number of NMR lines × relative intensity). The • corresponds to signal with double intensity. (b) Computational ¹³C NMR spectrum for Y₃N@D₂(35)-C₈₈. The experimental and calculated ¹³C shifts are provided in the SI. (c) Correlation between experimental and computational ¹³C NMR results.

As previously indicated, there are three isomers of empty-cage C₈₈ cage, C₂(7), C_s(17) and C₂(33), that are predicted to be thermodynamically and kinetically stable.^5,6 However, C₈₈ with D₂(35) symmetry is not included in this group. Computational calculations (DFT) summarized in Figure 3 suggest that the neutral IPR D₂(35)-C₈₈ has a small HOMO-LUMO gap (1.27 eV), indicating the lower stability of the neutral D₂(35)-C₈₈ cage. However, upon accepting six electrons, the HOMO-LUMO gap becomes significantly larger (1.77 eV), consistent for higher stability of the Y₃N@D₃(35)-C₈₈ molecule. To our knowledge, neither the Y₃N cluster or the D₂(35)-C₈₈ cage have been isolated, but when associated together, they form a stable Y₃N@D₃(35)-C₈₈ structure by electron transfer between the cluster and the cage.

Figure 3.

DFT computational HOMO-LUMO levels for the neutral IPR D₂(35)-C₈₈ and Y₃N@D₂(35)-C₈₈ cages.

As illustrated in Figure 4, the CV electrochemistry of Y₃N@C₈₈ and Gd₃N@C₈₈ are nearly equivalent. The Gd₃N@C₈₈ sample for this CV electrochemistry comparison was isolated and purified (see SI) in the same fashion as described for Y₃N@C₈₈ vide supra and is consistent with previously reported data by Echegoyen and coworkers.¹⁶ The electrochemistry of Y₃N@C₈₈ and Gd₃N@C₈₈ can be described as having two distinct oxidative processes accompanied by a first and second reduction peak potential of −1.43 V and −1.70V, respectively. Redox potentials and resulting electrochemical gap (ΔE_gap) for Y₃N@C₈₈ are in good agreement with reported literature values for other M₃N@C₈₈ systems (Table 1). This is consistent with a single (non-degenerate) LUMO level for Y₃N@C₈₈ obtained from the DFT calculations described above. The resulting electrochemical bandgap (^oxE₁-^redE₁) for the Y₃N@C₈₈ is 1.46 V, reasonably consistent with the DFT predictions above. Although (^oxE₁-^redE₁) for Y₃N@C₈₈ is smaller than the more stable Y₃N@C₈₀ case,¹⁶ this bandgap is similar to other M₃N@C₈₈ family members, such as, Gd₃N@C₈₈ (1.49 V),¹⁷ Nd₃N@C₈₈ (1.43 V),¹¹ Pr₃N@C₈₈ (1.43 V),⁹ Ce₃N@C₈₈(1.38 V)⁹ and La₃N@C₈₈(1.57 V)⁹. These results suggests that the (^oxE₁-^redE₁) bandgap (and corresponding HOMO-LUMO gap) of A₃N@C₈₈ trimetallic nitride endohedral metallofullerenes is a property dominated by the properties of the D₂(35)-C₈₈ carbon cage and not the nature of the yttrium or lanthanide electronic metal cluster (A₃N)⁶⁺.

Figure 4.

Cyclic voltammogram of Y₃N@C₈₈ and Gd₃N@C₈₈; 100 mV/s; 0.1 M TBABF₄ in o-DCB.

Table 1.

Redox potential of the A₃@C_2n Endohedral Metallofullerenes

	E1/2ox1	E1/2ox2	Epcred1	Epcred2	ΔEgap
Y3N@Ih-C8015	0.64		−1.41	−1.83	2.05
Y3N@C88	0.03	0.43	−1.43	−1.70	1.46
Gd3N@C88	0.05	0.45	−1.39	−1.71	1.44
Gd3N@C8816	0.06	0.49	−1.43	−1.74	1.49
Nd3N@C8810	0.07	0.53	−1.36	−1.75	1.43
Pr3N@C889	0.09	0.54	−1.34	−1.72	1.43
Ce3N@C889	0.08	0.63	−1.30	−1.57	1.38
La3N@C889	0.21	0.66	−1.34	−1.67	1.57

Conclusion

In summary, we have synthesized, purified and characterized diamagnetic Y₃N@C₈₈ for the first time. The ¹³C NMR study indicates that Y₃N@C₈₈ exhibits an IPR-obeying D₂(35)-C₈₈ cage. The electrochemical data suggested that Y₃N@C₈₈ has a smaller HOMO-LUMO gap than Y₃N@C₈₀ which is consistent with the computational DFT study. Also, our results show that encapsulation of an Y₃N cluster does not significantly alter the electrochemical properties of these trimetallic nitride endohedral metallofullerenes and this result is consistent with other lanthanide M₃N clusters in D₂(35)-C₈₈ cages. This suggests that the unique D₂(35)-C₈₈ cage properties strongly influence the electrochemical and electronic properties of these trimetallic nitride endohedral metallofullerenes.