Sc₃CH@C₈₀: selective ¹³C enrichment of the central carbon atom

Abstract

Sc₃CH@C₈₀ is synthesized and characterized by ¹H, ¹³C, and ⁴⁵Sc NMR. A large negative chemical shift of the proton, −11.73 ppm in the I_h and −8.79 ppm in the D_5h C₈₀ cage isomers, is found. ¹³C satellites in the ¹H NMR spectrum enabled indirect determination of the ¹³C chemical shift for the central carbon at 173 ± 1 ppm. Intensity of the satellites allowed determination of the ¹³C content for the central carbon atom. This unique possibility is applied to analyze the cluster/cage ¹³C distribution in mechanistic studies employing either ¹³CH₄ or ¹³C powder to enrich Sc₃CH@C₈₀ with ¹³C.

The pressure of helium gas is one of the most important parameters affecting the yield of fullerenes in arc-discharge synthesis. Optimization of the atmosphere in the arc-discharge generator (both the pressure and composition) is even more crucial for the synthesis of endohedral metallofullerenes (EMFs) and clusterfullerenes, whose yields are usually much lower than those of empty fullerenes.¹ New types of fullerenes or their derivatives can be obtained by introducing different reagents into the arc. Stevenson et al. were the first to show that in the presence of molecular nitrogen, the nitride clusterfullerene Sc₃N@C₈₀ can be synthesized in appreciable yield.² Then, Dunsch et al. demonstrated advantages of the reactive atmosphere method in the synthesis of EMFs: the use of NH₃ as a source of nitrogen not only afforded the synthesis of nitride clusterfullerenes, but also suppressed the yield of empty fullerenes.³ The method was adopted for the synthesis of other types of EMF clusterfullerenes, such as sulfides,⁴ oxides,⁵ or certain types of carbides.^6–9 It was also applied to stabilize unconventional empty fullerene cages via their in situ derivatization by hydrogen¹⁰ or chlorine atoms.¹¹ Interestingly, whereas the use of SO₂ or CO gases for the synthesis of clusterfullerenes leads to a large amount of empty fullerenes, the hydrogen-containing reagents (NH₃, CH₄, solid organic compounds¹²) suppress the yield of empty fullerenes.

Recently we have shown that methane can be advantageous for the synthesis of carbide clusterfullerenes, such as M₂TiC@C₈₀ or M₂TiC₂@C₈₀ (M is Y or a lanthanide).⁸ It is not clear if methane is barely a source of hydrogen suppressing the empty fullerene formation, or it plays a more specific role by, e.g., supplying the carbon for the endohedral cluster. Answering this question may shed more light on the fullerene formation, but the analysis of the carbon source in the EMF molecule is not straightforward. The use of isotopic substitution would be an obvious way to address this problem, but mass-spectrometry is not able to distinguish the cage and cluster atoms, whereas sensitivity of ¹³C NMR is not sufficient to allow distribution studies (detection of the central carbon atoms in carbide clusterfullerenes by ¹³C NMR required ¹³C enrichment^9,13–15). Here we circumvent this problem by studying the Sc₃CH@C₈₀, which affords ¹³C analysis of the central carbon via the ¹H NMR signal of the endohedral hydrogen and show that methane plays an active role in the formation of the endohedral cluster.

Sc₃CH@C₈₀ was synthesized in two series of experiments, using either pure Sc or a 1 : 1 mixture of Sc and Ti, as a source of metal. Metals were mixed with graphite powder and packed into the hole-drilled graphite rods, which were then used in the arc-discharge synthesis in the He atmosphere with an addition of several mbar of CH₄ (250 mbar total pressure). The main EMF products of the arc discharge synthesis in these conditions are Sc₄C₂@C₈₀ in pure Sc system, and Sc₂TiC@C₈₀ in the mixed-metal Sc/Ti system. Both systems afforded appreciable amounts of Sc₃CH@C₈₀, which was further isolated using HPLC (see ESI † for further details of separation).

Detailed characterization of Sc₃CH@C₈₀ in the first report on its synthesis was not possible due to the tiny amounts of the isolated compound.⁶ In this work, we accomplished characterization of the compound by ⁴⁵Sc, ¹³C, and ¹H NMR spectroscopy as shown in Fig. 1. The icosahedral cage symmetry of Sc₃CH@C₈₀ is proved by ¹³C NMR spectroscopy (Fig. 1a). Two cage resonances with the 3 : 1 intensity ratio are observed at 144.06 and 136.78 ppm, which is close to the chemical shifts reported for other clusterfullerenes with the C₈₀-I_h cage^2,16–19 (144.57/137.24 ppm in Sc₃N@C₈₀,² 144.7/137.8 in Sc₄C₂@C₈₀,¹⁷ 144.9/137.7 ppm in Sc₃CN@C₈₀,¹⁸ or 144.82/137.29 ppm in Sc₄O₂@C₈₀¹⁹). In the ¹³C-enriched sample, satellite peaks due to coupling of neighboring cage atoms can be seen (Fig. 1a) with the ¹J_CC coupling constant of 58 Hz, typical for C-sp² carbon atoms in conjugated π-systems.²⁰ Presumably, the large line-width of the endohedral carbon signal¹⁴ did not allow us to detect it in the direct ¹³C NMR measurements, and the chemical shift of the endohedral carbon was determined from selective decoupling measurements of ¹H NMR (see below).

Fig. 1.

NMR spectra of Sc₃CH@C₈₀ dissolved in CS₂: (a) 125 MHz ¹³C NMR; (b) 121.5 MHz ⁴⁵Sc NMR (black line – Sc₃CH@C₈₀, cyan line – Sc₃N@C₈₀); (c) 500 MHz ¹H NMR. The insets in (a) and (c) show ¹³C satellites marked with asterisks.

In ⁴⁵Sc NMR spectrum, Sc₃CH@C₈₀ exhibits a single resonance at 292 ppm (Fig. 1b). This value is 100 ppm low-field with respect to the ⁴⁵Sc chemical shift in Sc₃N@C₈₀ at 191 ppm (Fig. 1b). The proton NMR signal of Sc₃CH@C₈₀ was detected at −11.73 ppm (Fig. 1c). Such a high-field value is typical for endohedral protons^21–24 (see Table 1). We also detected formation of the second isomer of Sc₃CH@C₈₀, presumably with the D_5h carbon cage (Fig. S6, ESI†), whose ¹H NMR signal is detected at −8.79 ppm.

Table 1.

¹H and ¹³C chemical shifts (ppm) for endohedral clusters in Sc₃CH@C₈₀ and selected endohedral fullerenes

EMF	δ(1H)	Ref.	EMF	δ(13C)	Ref.
Sc3CH@C80-Ia	−11.73		Sc3CH@C80-Ia	173 ± 1
Sc3CH@C80-IIa	−8.79		M2C2@C2nb	220–260	13–15 and 25
H2@C60	−1.44	21	YCN@C82	292.4	26
H2O@C60	−4.81	22	Sc3C2@C80−	328.3	14
H2@C70	−23.97	23	Lu2TiC@C80	340.98	9

^aSc₃CH@C₈₀-I and Sc₃CH@C₈₀-II denote the major (I_h) and the minor (presumably D_5h) isomers.

^bM = Sc, Y; 2n = 80, 82, 84, 92.

The ¹H NMR spectrum of Sc₃CH@C₈₀ also exhibits a low intensity doublet due to the proton-bonded ¹³C (Fig. 1c). The ¹J_C–H coupling constant is small, 78.5 Hz, which is typical for protons bonded to carbon atoms with highly electropositive substituents. Thus, the measured ¹J_C–H constant is in line with the large negative charge on the central carbon atom predicted for Sc₃CH@C₈₀.²⁷ The possibility to detect ¹³C satellites in the proton NMR spectrum enables determination of the ¹³C chemical shift of the central carbon atom via ¹H NMR measurements with selective ¹³C decoupling at different ¹³C irradiation frequencies. The satellites are well visible at 165 or 180 ppm, but disappear completely at 172–175 ppm (Fig. S4, ESI†). Thus, the ¹³C chemical shift of the central atom in Sc₃CH@C₈₀ is determined as 173 ± 1 ppm, which is noticeably downfield than ¹³C chemical shifts of endohedral carbons in carbide clusterfullerenes (see Table 1).

Sc₃CH and Sc₃N clusters have the same formal charge (6+) and identical electron count (the (C–H)³⁻ unit is analogous to the nitride ion N³⁻), and therefore a close similarity of the electronic properties of Sc₃CH@C₈₀ and Sc₃N@C₈₀ can be expected.^6,27 Indeed, both compounds exhibit very similar Vis-NIR absorption spectra (Fig. 2a), with a slight blue shift of the lowest energy band in Sc₃CH@C₈₀ (717 nm versus 734 nm in Sc₃N@C₈₀). The difference is more distinct in the fluorescence spectra: whereas Sc₃CH@C₈₀ exhibits a NIR emission band at 835 nm, the maximum of the fluorescence band of Sc₃N@C₈₀ is observed at 910 nm. Crossing points of the absorption and emission bands give the optical gaps of Sc₃CH@C₈₀ and Sc₃N@C₈₀ as ca. 1.62 and 1.53 eV, respectively. The electrochemical gap of Sc₃CH@C₈₀ is also 0.09 V larger than that of Sc₃N@C₈₀ (Fig. 2b and Table 2). Both compounds exhibit similar redox behaviour with chemically irreversible first reduction near −1.2 V. DFT calculations show that Sc₃CH@C₈₀ and Sc₃N@C₈₀ have almost identical spatial distribution of the HOMO and LUMO. The HOMO is essentially a carbon cage orbital, whereas the LUMO has large contribution of Sc atoms (Fig. 2c).

Fig. 2.

(a) UV-Vis spectra of Sc₃CH@C₈₀ (black) and Sc₃N@C₈₀ (cyan) in toluene, the inset shows absorption spectra in the range of the lowest energy transitions and luminescence spectra (Sc₃CH@C₈₀ – red, Sc₃N@C₈₀ – magenta; laser excitation at λ_ex = 405 nm); (b) square wave voltammetry of Sc₃CH@C₈₀ (black) and Sc₃N@C₈₀ (cyan) in o-dichloro-benzene/TBABF₄, asterisks mark Fe(Cp)₂ and Fe(Cp*)₂ used as internal standards; to guide an eye, the first reduction and oxidation potentials of Sc₃CH@C₈₀ are denoted with vertical red lines; (c) HOMO and LUMO of Sc₃CH@C₈₀ computed at the PBE/def2-TZVP level.²⁸

Table 2.

Redox potentials (V) of Sc₃CH@C₈₀ and Sc₃N@C₈₀^a

EMF	O-II	O-I	R-I	R-II	R-III	GapEC
Sc3CH@C80		0.67	−1.21	−1.53/−1.82	−2.28	1.88
Sc3N@C80	1.09	0.63	−1.15	−1.54/−1.73		1.79

^aAll potentials are determined by square-wave voltammetry in o-dichloro-benzene/TBABF₄ and are referred versus Fe(Cp)₂^+/0 redox couple; “O” and “R” denote oxidation and reduction, respectively.

Sc₃CH@C₈₀ offers a unique possibility to study the role of methane in the carbide clusterfullerene formation using ¹³C-enrichment. The isotopic distribution of the central carbon atom can be determined by ¹H NMR from the intensity of the ¹³C satellites, whereas the net isotopic distribution in the whole molecule (dominated by that of the carbon cage) can be deduced from the mass-spectra. We synthesized ¹³C-enriched Sc₃CH@C₈₀ by applying either (i) ¹³CH₄ or (ii) ¹³C powder. To distinguish the two series, they will be denoted as “¹³CH₄/C” and “CH₄/¹³C”, respectively. The amount of ¹³C powder in the CH₄/¹³C series was adjusted to keep the same amount of ¹³C in the generator as in the ¹³CH₄/C series. Fig. 3 compares ¹H NMR and mass-spectra of the Sc₃CH@C₈₀ sample synthesized with the natural-abundant CH₄/C to the two types of ¹³C-enriched samples, whereas estimated ¹³C content is summarized in Table 3. In the CH₄/¹³C syntheses, the isotopic composition for the central atom and the whole molecule are both equal 5–6% within the uncertainty limits of the NMR measurements (ca. 1%). Thus, the carbon originating from the powder is equally distributed between the cage and the central atom; similar conclusion was achieved by Dorn et al. in their mass-spectrometric study of the empty fullerenes and Y–carbide clusterfullerenes.¹⁵

Fig. 3.

(a) Mass-spectra of Sc₃CH@C₈₀ samples with different ¹³C content obtained in CH₄/¹³C, ¹³CH₄/C, and CH₄/C syntheses; (b) ¹H NMR spectra for the same samples, normalized to the intensity of the main singlet.

Table 3.

¹³C content for the central atom and the whole molecule

13C enrichment	1H NMR (%)	Mass-spectrometry (%)
C/CH4a	1.1 ± 0.4	1.1 ± 0.1
CH4/13C	5.8 ± 0.9	5.0 ± 0.2
13CH4/C	7.6 ± 1.5	1.6 ± 0.1

^aNatural abundance.

Substantially different results are obtained in the ¹³CH₄/C syntheses: the ¹³C enrichment of the carbon cage is only 1.6 ± 0.1%, whereas the ¹³C content for the central atoms is much higher, 7.6 ± 1.5%. Thus, despite the rather large error bars in the NMR measurement (caused by the limited sample amount), selective ¹³C enrichment of the central carbon atom by ¹³CH₄ is beyond any doubt. The ¹³CH₄/C syntheses thus provide rich information on the Sc₃CH@C₈₀ formation process.

The volume inside the generator can be schematically divided into three zones:

• (1)The “hot” zone near the center of the arc, where the temperature is up to several thousand K,²⁹ and majority of chemical bonds (including C–H) are broken. Only the most stable species (such as C₂ dimers) can survive.
• (2)The periphery of the hot zone, where the carbon vapor cools down by the adiabatic expansion and interaction with helium atoms, resulting in a self-assembly of fullerenes and other carbonaceous structures. This intermediate zone is hot enough to provide sufficient energy for the rearrangement of the building carbon networks, but the temperature is not high enough for their atomization.
• (3)The “cold” zone, where the fullerenes can only anneal (e.g., structural defects can be healed), but substantial structural rearrangements are already not possible.

We propose that all CH₄ molecules entering the hot zone are completely atomized, and therefore the carbon atoms from methane can serve as a source of carbon for the fullerene cage. Since the hot zone occupies only a small volume, whereas methane is distributed over the whole generator chamber, only a small fraction of methane present in the system passes through the hot zone. Thus, it is not surprising that the content of methane-originating carbon atoms in the fullerene cage is not exceeding 0.5%, whereas the main sources of carbon for both the fullerene cage and the endohedral cluster are the graphite rods and graphite powder packed into the rods.

When the C/H vapor leaves the hot zone, the CH bonds can be formed again. Note that from the 0.5% contribution of the methane as a source of carbon for the cage, the C : H ratio in the hot zone is tentatively estimated as 50 : 1. This ratio is sufficient for a dramatic suppression of the empty fullerene formation. The ¹³C/¹²C distribution in the newly formed CH bonds can be considered roughly equal the ¹³C content in the carbon cage (1.6% in the ¹³CH₄/C syntheses). The fact that the ¹³C content for the central carbon atom in Sc₃CH@C₈₀ obtained in the ¹³CH₄ syntheses is several times higher than for the fullerene cage means that methane is also chemically active in the “intermediate” zone. Although CH₄ molecules are not completely atomized here, they can exchange protons with other carbon structures or react with Sc atoms to substitute protons. However, some ¹³CH fragments remain intact (else the isotopic distribution for the cage and the cluster would be equalized) and take part in the endohedral fullerene formation. The fraction of such “native” ¹³CH units in Sc₃CH@C₈₀ is the difference between the ¹³C content in the cluster and in the cage, and can be roughly estimated as 6%.

In this work we reported on the synthesis and spectroscopic characterization of Sc₃CH@C₈₀. Its ¹³C, ⁴⁵Sc, and ¹H NMR spectra are reported for the first time and fully establish the molecular structure of this clusterfullerene. Electronic properties of Sc₃CH@C₈₀ are similar to those of its close analog, nitride clusterfullerene Sc₃N@C₈₀. Yet, absorption and fluorescence spectroscopy as well as electrochemical study show that the band-gap of Sc₃CH@C₈₀ is higher by 0.09 eV. Most importantly, a unique possibility to determine ¹³C composition of the central atom in the cluster by ¹H NMR enables an analysis of the role of methane in the clusterfullerene formation. A series of ¹³C enrichment with either ¹³CH₄ or ¹³C powder showed that the use of ¹³CH₄ in the synthesis of Sc₃CH@C₈₀ allows selective enrichment of the central carbon atom with ¹³C.

Supplementary Material

†Electronic supplementary information (ESI) available: Additional experimental details, HPLC separation, FTIR and NMR spectra. See DOI: 10.1039/c5cc10025a