C ₂ C₇₀F₃₈ is aromatic, contains three planar hexagons, and has equatorial addends

Abstract

Single crystal X-ray analysis shows the main (C₂) isomer of C₇₀F₃₈ to contain three planar delocalised aromatic hexagons (two equivalent and one centred on the C₂ axis), together with seven C=C bonds (three pairs and one straddling the C₂ axis); C₇₀F₃₈ is the first high addition level [70]fullerene derivative to be fully characterised, is the first to have equatorial addends, and is calculated to have high stability.

---

Understanding the addition patterns of fullerenes and their origins is an integral part of fullerene chemistry. For [60]fullerene, numerous papers describe not only numerous fully characterised cycloadducts, but also the structures of derivatives (especially those containing H, F, Cl, Br) at a range of addition levels, enabling conjectures of the stepwise addition pathways.¹

The situation for [70]fullerene is more complex,² and involves two main manifolds: additions/cycloadditions to the end caps giving e.g. C₇₀H_2/4, C₇₀Me₂, C₇₀benzyl₂, C₇₀(OH)₂, C₇₀O, and zig-zag equatorial addition giving C₇₀Me_4/6/8/10, C₇₀Ph_2/4/6/8/10, C₇₀H_8/10, C₇₀Cl₁₀, C₇₀Br₁₀, C₇₀(O₂t-Bu)₁₀, C₇₀Ph₉OH, C₇₀Ph₈(OH)₂. Low addition-level products have also been obtained in reactions with benzyne, bromomalonate, Vaska's complex, and 2,2-bis(2,6-diisopropylphenyl)hexamethyltrisilane. In no reaction has equatorial addition ever been observed. (Reported equatorial silylation of [70]fullerene was based on ¹H NMR data, consistent however with equatorial straddling).³

Information on high levels of [70]fullerene radical-addition is sparse, and these have been obtained only in hydrogenation (which gave mainly uncharacterised C₇₀H_36/38 together with some C₇₀H_40–44),⁴ and fluorination by MnF₃ (in which we isolated twenty-one highly fluorinated [70]fullerenes (mainly C₇₀F_36/38 with some C₇₀F_34–42), and many oxide derivatives.⁵ However, the very complex ¹⁹F NMR spectra precluded structural identification.

A feature of these additions is the absence of intermediate addition levels and the dominance of C₇₀X₃₆ and C₇₀X₃₈ (X = H, F). This led us to propose that C₇₀H₃₆ must be an aromatic compound, having benzenoid patches in the structure.⁶ Aromatic structures were also proposed by Gerst et al., and by Book and Scuseria.⁷ However, subsequent MNDO calculations indicated that a structure having isolated double bonds would, for steric reasons, be much more stable than an aromatic one.⁸ Clare and Kepert have calculated structures for C₇₀X_36/38 (X = H, F)⁹ and predicted a number of structures for C₇₀F₃₈, including some having three aromatic rings and seven isolated double bonds. Since the 36 and 38 addition levels are clearly closely related, our target was to obtain a crystal structure to resolve both the aromaticity question and subsequently to deduce the formation route. We have now succeeded in our objective, and provide the first full characterisation of a highly addended [70]fullerene.

Fluorination of [70]fullerene by different fluorinating reagents was explored briefly, using a 5–10-fold weight excess of fluorinating reagent at ca. 520 °C; [70]fullerene is much less reactive than [60]fullerene. Each of BaPbF₆,¹⁰ MnF₃, and CeF₄ showed C₇₀F₃₈ as the main product, with BaPbF₆ giving the highest relative yield. CeF₄ gave only C₇₀F₃₆ and C₇₀F₃₈, and very poor overall yields together with some unreacted fullerene. The other two reagents showed the presence of C₇₀F₄₀, whilst CoF₃ gave significant quantities of C₇₀F_40–50. Because of the extreme demands of the HPLC separations,⁵ the products were combined and processed together. Numerous derivatives were obtained as before,⁵ and also some new ones (due probably to the different fluorinating reagents, and the larger scale); this will be described in a full paper. As before, the main C₇₀F₃₈ isomer was no. 14,⁵ which crystallised from toluene.

The single-crystal X-ray structure† (Figs. 1 and 2), reveals the structure to be aromatic, having three planar delocalised benzenoid rings, together with seven isolated double bonds. Their location is most easily seen from the Schlegel diagram (Fig. 3), which also gives the numbering. The view from the end caps (Fig. 2) shows a ‘triangular’ shape imparted by the presence of the three planar hexagons. Two of the benzenoid rings (Ar) are equivalent, whilst the third (Ar^*) is unique and sits on the (non-crystallographic) C₂ axis. The double bonds comprise three equivalent pairs (C4–5/C40–41; C12–13/C68–69; C15–16/C43–60), and the unique C20–21 straddling the C₂ axis. A unique structural feature is the presence of two equatorial fluorines; equatorial addends on [70]fullerene have never been seen before.

Fig. 1 ORTEP drawing (20% ellipsoids) for C₂ C₇₀F₃₈, viewed down the C₂ axis; the facing planar central hexagon is Ar^* and the rear central C=C bond is C20–C21 (see Fig. 3).

Fig. 2 ORTEP drawing (20% ellipsoids) for C₂ C₇₀F₃₈ viewed from an end cap, showing the ‘triangular’ shape imparted by the planar regions.

Fig. 3 Schlegel diagram for C₂ C₇₀F₃₈ showing the aromatic rings and the isolated double bonds (• = F). The C₂ axis bisects the C20–C21 bond and passes through the centre of the aromatic ring Ar^*.

The bond lengths (Table 1) are the average values for symmetry equivalent bonds given in terms of the lowest locant numbers. The average bond lengths (Å) in the three aromatic rings are 1.381(5), 1.384(5) and 1.382(5), confirming their aromaticity, cf. the values found in C₁ C₆₀F₃₆ (1.371, 1.378, 1.389),¹¹T C₆₀F₃₆¹² and C₆₀F₁₈¹³ (both 1.373). The C15=C16/C43=C60 double bonds are exceptionally short (1.313 Å average), attributed to compression produced by the neighbouring fluorines (cf. the short isolated double bonds in both C₁ C₆₀F₃₆¹¹ and C₆₀F₄₈).¹⁴ As with C₁ C₆₀F₃₆ there is an exceptionally long FC–CF bond of 1.642 Å (27,28; 52,66), which arises as in the previous example, where a carbon is surrounded by three CF groups. The planarity of the benzenoid rings is shown by the valence angle sums for the aromatic carbons, which average 359.0° for Ar/Ar and 359.6° for Ar^*.

Table 1 Cage C–C bond lengths (Å)^a for C₂ C₇₀F₃₈

Bond		Bond		Bond
a These are the average values for symmetry equivalent bonds, and given as lowest locants; esd values are all 0.013 ± 0.001 Å.
C1–2	1.592	C10–C26	1.377	C25–C44	1.499
C1–6	1.580	C11–C12	1.507	C26–C27	1.505
C1–9	1.498	C11–C28	1.591	C27–C28	1.642
C2–3	1.555	C12–C13	1.328	C27–C46	1.547
C2–12	1.543	C13–C14	1.498	C28–C29	1.536
C3–4	1.509	C13–C30	1.478	C29–C30	1.585
C3–14	1.568	C14–C15	1.501	C29–C48	1.481
C4–5	1.343	C15–C16	1.313	C30–C31	1.565
C4–17	1.498	C15–C32	1.508	C31–C32	1.557
C5–6	1.470	C16–C17	1.492	C31–C49	1.496
C5–19	1.498	C16–C34	1.480	C32–C33	1.615
C6–7	1.574	C17–C18	1.552	C33–C34	1.587
C7–8	1.508	C18–C19	1.578	C33–C51	1.601
C7–21	1.500	C18–C36	1.510	C46–C47	1.481
C8–9	1.378	C19–C20	1.546	C47–C48	1.382
C8–24	1.398	C20–C21	1.335	C47–C63	1.375
C9–10	1.355	C24–C25	1.392	C48–C49	1.391
C10–11	1.478	C25–C26	1.395

---

Clare and Kepert proposed nineteen possible structures for C₇₀F₃₈ based on AM1 calculations, five of which had three aromatic rings and seven isolated double bonds,⁹ though none correspond to the characterised structure. Previously, we isolated seven other isomers of C₇₀F₃₈ (all C₁) and it seems probable that these arise merely from 1,3-migrations of fluorines giving alternative locations of the double bonds (there are fourteen possibilities), producing C₁ structures. We hope to investigate this aspect further. We have calculated (AM1) the heat of formation of C₂ C₇₀F₃₈ and find it to be ca. 24 and 19 kcal mol⁻¹ lower than the values for the most stable structures (nos. 1 and 2, respectively) in Table 4 of the Clare and Kepert paper.

The other highly fluorinated species C₇₀F_34/36 and C_40/50 are probably also aromatic, with more or fewer double bonds respectively, but retaining the locations of the benzenoid hexagons. The present structure thus presents a basis for extensive further calculations.

The ¹⁹F NMR spectrum that we obtained previously for this isomer showed nineteen lines (confirmed here), two of which were upfield multiplets at δ −164.59 and −168.19. Our previous work with fluorinated fullerenes showed that upfield multiplets arise from fluorines attached to carbons having three CF neighbours, and are therefore assigned to fluorines at C28/C66, and C33/C45, which are c- and d-type¹⁵ carbons, respectively. Since c-type carbons appear more downfield, we provisionally assign the δ −164.59 peak to F28/66. We hope to be able to assign all of the NMR peaks, and hence establish the peak positions of fluorines in specific locations in C₇₀F_n, and thereby solve all of the other structures.

Lastly, C₂ C₇₀F₃₈ does not contain the motif present in C₇₀Cl₁₀etc., suggesting that addition commences at the end caps, though isolation of intermediate structures would be necessary to support this conjecture.

Notes and references

1. R. Taylor, Lecture Notes on Fullerene Chemistry: A Handbook for Chemists, Imperial College Press, London, 1999, ch. 4, 7, 9, 11.
2. Leading references are collated in: R. Taylor, J. Fluorine Chem., 2004, 125, 359.
3. T. Akasaka, E. Mitsihida, W. Ando, K. Kobayashi and S. Nagau, J. Am. Chem. Soc., 1994, 116, 2627.
4. A. D. Darwish, A. K. Abdul-Sada, G. J. Langley, H. W. Kroto, R. Taylor and D. R. M. Walton, J. Chem. Soc., Perkin Trans. 2, 1995, 2359.
5. R. Taylor, A. K. Abdul-Sada, O. V. Boltalina and J. M. Street, J. Chem. Soc., Perkin Trans. 2, 2000, 1013.
6. R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1994, 2497.
7. M. Gerst, H.-D. Beckhaus, C. Rüchardt, E. E. B. Campbell and R. Tellgman, Tetrahedron Lett., 1993, 34, 7729; L. D. Book and G. E. Scuseria, J. Phys. Chem., 1994, 98, 4283.
8. P. W. Fowler, J. P. B. Sandall, S. J. Austin, D. E. Manolopoulos, P. D. M. Lawrenson and J. M. Smallwood, Synth. Met., 1006, 77, 97; P. W. Fowler, J. P. B. Sandall and S. J. Austin, Fullerene Sci. Technol., 1996, 4, 369; P. W. Fowler, J. P. B. Sandall and R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1997, 419.
9. B. W. Clare and D. E. Kepert, J. Mol. Struct.: THEOCHEM, 2002, 583, 19–45.
10. P. A. Troshin, O. V. Boltalina, N. V. Polyakova and Z. E. Klinkina, J. Fluorine Chem., 2001, 110, 157.
11. A. G. Avent, B. W. Clare, P. B. Hitchcock, D. L. Kepert and R. Taylor, Chem. Commun., 2002, 2370.
12. P. B. Hitchcock and R. Taylor, Chem. Commun., 2002, 2078.
13. I. S. Neretin, K. A. Lyssenko, M. Yu. Antipin, Yu. L. Slovokhotov, O. V. Boltalina, P. A. Troshin, A. Yu. Lukonin, L. N. Sidorov and R. Taylor, Angew. Chem., Int. Ed., 2000, 39, 3273.
14. S. I. Troyanov, P. A. Troshin, O. V. Boltalina, I. N. Ioffe, L. N. Sidorov and E. Kemnitz, Angew. Chem., Int. Ed., 2001, 40, 2285.
15. R. Taylor, J. P. Hare, A. K. Abdul-Sada and H. W. Kroto, J. Chem. Soc., Chem. Commun., 1990, 1423.

---

Footnote

† Crystal data: C₇₀F₃₈·2.5(C₇H₈), M = 1793.0, triclinic, space group P1̄ (no. 2), a = 12.1508(6), b = 15.5160(7), c = 18.0125(8) Å, α = 94.308(3), β = 101.526(2), γ = 106.457(2)°, V = 3159.6 ų, Z = 2, D_x = 1.89 Mg m⁻³, μ(Mo-Kα) = 0.19 mm⁻¹, T = 173 K. Diffraction from the crystal was weak and limited. 11057 unique reflections (R_int = 0.13) measured on a KappaCCD diffractometer. Three disordered toluene solvate molecules included with restrained geometry. Refinement on F² using SHELX97, R1 = 0.110 for 5490 reflections with I > 2σ(I), wR2 = 0.324 for all reflections. CCDC 247743. See http://www.rsc.org/suppdata/cc/b4/b412599a/ for crystallographic data in .cif or other electronic format.

---