Chlorination-Promoted Skeletal Transformations of Fullerenes

Jun 10, 2019 - To date, these include single- and multistep SWRs in the buckminsterfullerene C60 and in the higher fullerenes C76(1), C78(2), C82(3), ...
0 downloads 0 Views 8MB Size
Download PDF
Article Cite This: Acc. Chem. Res. 2019, 52, 1783−1792

pubs.acs.org/accounts

Chlorination-Promoted Skeletal Transformations of Fullerenes Published as part of the Accounts of Chemical Research special issue “Advanced Molecular Nanocarbons”. Shangfeng Yang,*,† Ilya N. Ioffe,‡ and Sergey I. Troyanov*,‡ †

Downloaded via KEAN UNIV on July 17, 2019 at 07:44:48 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Hefei National Laboratory for Physical Sciences at Microscale, CAS Key Laboratory of Materials for Energy Conversion, Department of Materials Science and Engineering, University of Science and Technology of China, Hefei 230026, China ‡ Department of Chemistry, Moscow State University, 119991 Moscow, Russia CONSPECTUS: Classical fullerenes are built of pentagonal and hexagonal rings, and the conventional syntheses produce only those isomers that obey the isolated-pentagon rule (IPR), where all pentagonal rings are separated from each other by hexagonal rings. Upon exohedral derivatization, the IPR fullerene cages normally retain their connectivity pattern. However, it has been discovered that high-temperature chlorination of fullerenes with SbCl5 or VCl4 can induce skeletal transformations that alter the carbon cage topology, as directly evidenced by single crystal X-ray diffraction studies of the chlorinated products of a series of fullerenes in the broad range of C60 to C102. Two general types of transformations have been identified: (i) the Stone−Wales rearrangement (SWR) that consists of a rotation of a C−C bond by 90°, and (ii) the removal of a C−C bond, i.e., C2 loss (C2L). Single- or multistep SWR and/or C2L transformations afford either classical non-IPR fullerenes bearing fused pentagons (highlighted in red in the TOC picture) or nonclassical (NCx) fullerenes with x = 1−3 heptagonal rings (highlighted in blue in the TOC picture) often flanked by fused pentagons. Several subtypes of the SWR and C2L processes can be further discerned depending on the local topology of the transformed region of the cage. Under the chlorination conditions, the non-IPR and NC carbon cages that would be energetically unfavorable and mostly labile in their pristine state are instantaneously stabilized by chlorination of the pentagon-pentagon junctions and by delimitation of the original spherical π-system into smaller favorable aromatic fragments. The significance of the chlorination-promoted skeletal transformations within the realm of fullerene chemistry is demonstrated by the growing body of examples. To date, these include single- and multistep SWRs in the buckminsterfullerene C60 and in the higher fullerenes C76(1), C78(2), C82(3), and C102(19), single and multistep C2Ls (i.e., cage shrinkage) in C86(16), C88(33), C90(28), C92(50), C96(80), C96(114), and C102(19), and multistep combinations of SWRs and C2Ls in C88(3), C88(33), and C100(18), (IPR isomer numbering in parentheses is according to the spiral algorithm). Remarkably, an IPR precursor can give rise to versatile transformed chlorinated fullerene cages formed via branched pathways. The products can be recovered either in their initial chlorinated form or as more soluble CF3/F derivatives obtained by an additional trifluoromethylation workup. Reconstruction of the skeletal transformation pathways is often complicated due to the lack of the isolable intermediate products in the multistep cases. Therefore, it is usually based on the principle of selecting the shortest pathways between the starting and the final cage. The quantum-chemical calculations illustrate the detailed mechanisms of the SWR and C2L transformations and the thermodynamic driving forces behind them. A particularly important aspect is the interplay between the chlorination patterns and the regiochemistry of the skeletal transformations. rearrangement (SWR),3 consists in a 90° rotation of the central C−C bond in a pyracylene patch and results in alteration of the connectivity pattern (Figure 1a). However, pristine fullerenes are stable against SWRs because of considerable activation barriers4 and, except for the IPR-toIPR cases, high endothermic effect. Therefore, the developments in the chemistry of pristine fullerenes in the past nearly three decades have been mostly associated with the exohedral derivatization without any skeletal transformations of the carbon cages.5 However, an earlier theoretical work pointed out that the increased binding energy of addends at the pentagon−

1. INTRODUCTION Classical fullerene molecules are built of pentagonal and hexagonal rings, and the presence of 12 pentagons is required to form a closed polyhedron. Due to the closed shape of the molecules, their sp2 carbon atoms are slightly pyramidalized, which is energetically unfavorable. In the structures with fused pentagons, one would expect enhanced local pyramidalization together with the increased reactivity. Therefore, conventionally synthesized pristine fullerenes obey the isolated-pentagon rule (IPR);1 i.e., their pentagons are separated from each other by hexagons. Pristine IPR fullerenes are very stable against skeletal rearrangements. Pyrolytic degradation of larger fullerenes into smaller ones occurs only above 1000 °C.2 A possible nondestructive isomerization mechanism, a Stone−Wales © 2019 American Chemical Society

Received: April 10, 2019 Published: June 10, 2019 1783

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research

temperature (450 °C) trifluoromethylation of the chlorinated products with CF3I. It affords far better soluble trifluoromethylated derivatives recoverable by HPLC separation and crystallization from solution, as evidenced by the successes on the identifications of several new non-IPR and nonclassical derivatives of C6018 and C76.41 Trifluoromethylation itself has never been observed to promote skeletal transformations of fullerenes,42 so it only helps to capture those fullerenes that form in the course of chlorination.

Figure 1. Two basic skeletal transformations observed in fullerenes. (a) Stone−Wales rearrangements (SWRs) typically occur in pyracylene patch, resulting in an interchange of the positions of its pentagons and hexagons. (b) C2 loss (C2L) usually involves the removal of a 5:6 C−C bond, creating a heptagon and an additional pentagon in the carbon cage.

2.2. Topological Reconstruction and Theoretical Calculations

In the multistep skeletal transformations of fullerenes, the intermediate compounds are often elusive, and even the starting higher fullerene mixtures are, sometimes, characterized only mass spectrometrically. In those cases, skeletal transformation pathways need to be reconstructed on the basis of topological analysis. The reconstruction is based on the empirical rule of preferability of the shortest pathways from the known or MS-compliant starting cage to the final one. Obviously, the longer pathways will be less probable because of generally lower per-step driving force. However, the shortest pathways are often nonunique, and the most probable pathway can be identified only with the aid of a detailed quantumchemical analysis. To rationalize the basic transformation mechanisms and provide quantitative estimates of their energetics, we performed density functional theory (DFT; PBE/TZ2P) calculations of the reaction profiles with the use of the PRIRODA software.43,44 The results obtained with the PBE functional were also checked against the PBE0/Def2-SVP values with D-PCM corrections for the effects of the chlorinating medium calculated using the Firefly software partly based on the GAMESS(US) source code.45 It was found that the gas-phase PBE results agree well with the PCMcorrected PBE0 values and thus can be used as reliable estimates of the reaction barriers and heats.

pentagon junctions can stabilize derivatized non-IPR fullerenes with respect to their IPR counterparts.6 Indeed, a number of such non-IPR fullerene derivatives have been isolated via the dopant-modified arc-discharge or combustion methods.7−11 Chemical derivatization, such as halogenation, under harsh conditions is also a potential route to destructive cage transformations. As the first report, fluorination of C60 with cesium−lead oxyfluorides at 550 °C resulted in a C2 loss with reconnection of the adjacent bonds (hereafter abbreviated as C2L; see Figure 1b), yielding C58F18 and C58F17(CF3) byproducts where the C58 cage contains a heptagon and fused pentagons.12 A common approach to the synthesis of the derivatized unconventional fullerenes from the IPR precursors was discovered in 2009−2010. Both the SWR and C2L processes were found to occur under prolonged high-temperature chlorination of higher fullerenes.13,14 In the present Account, we summarize the results obtained in this field since 2009: the library of single- or multistep transformations in the IPR cages that give various classical (i.e., pentagon-hexagon) non-IPR fullerene derivatives and nonclassical compounds with one to three heptagons.

2. SYNTHESIS AND RESEARCH METHODOLOGY 2.1. Synthesis and Identification

The synthesis of the unconventional carbon cages is based on the high-temperature (340−360 °C) chlorination of the IPR fullerenes precursors with an excess of inorganic chlorides: SbCl5, VCl4, or their mixtures. Note that, at lower temperatures, only conventional IPR fullerene chlorides are obtained without skeletal changes. For the most abundant fullerenes, C60 and C70, chlorination below 400 °C still afforded only the corresponding IPR chlorides,16,17 but recently C60 was found to rearrange at higher temperatures of 420−440 °C.18 The reaction is typically carried out in thick-walled glass ampules for a period from several days to several months. It was found that SbCl5 more readily promotes the SWRs while VCl4 often leads to the C2L acts. The indirect data of various spectroscopic techniques are typically not helpful for the purposes of structure elucidation of the transformed products. Therefore, the structural identification is carried out by single crystal X-ray diffraction with the use of synchrotron radiation. Apart from the non-IPR and nonclassical compounds in the range of C60 to C102,19−30 the X-ray studies also revealed a number of new IPR fullerenes between C82 and C108.31−40 An obvious limitation of the XRD studies is a need for crystalline material of appropriate quality. There is always a possibility that some products form an amorphous phase or too fine crystals and thus get missed. This shortcoming stimulated a recent development that consists of in high-

3. A SURVEY OF SKELETAL TRANSFORMATIONS OF FULLERENES 3.1. Nomenclature

Hereafter, numbering of the classical fullerenes is based on the Fowler−Manolopoulos spiral algorithm,15 and the isomer number is given in superscript before the formula (e.g., 39,712 C82Cl28). The IPR fullerene can be numbered separately according to their order in the full list, and the number is then given in parentheses after the formula (e.g., C82(3) is another designation of 39,712C82). For the heptagon-containing nonclassical (NC) fullerenes we presently use nonunique tentative designations that refer to the number of the heptagonal rings in the cage (e.g., C60(NC1) or C96(NC3)). 3.2. Skeletal SWR Transformations

Although SWR transformations in C6018 are among the most recently discovered cases, we start our review from this basic and the most abundant fullerene. The structurally characterized transformation products formed via one to five SWR steps are shown in Figure 2. Due to the icosahedral symmetry of the buckminsterfullerene all its pyracylene patches are equivalent, and the first step is hence unique, insofar as the addition patterns are considered, giving 1809C60. Note that the 1809 C60 and 1804C60 cages were previously obtained through dopant-modified fullerene syntheses.9,10 In our work, it was 1784

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research

The above terms of SWR1 and SWR2 refer to the topological classification of the skeletal processes shown in Figure 3. While SWR1 is a classical SWR in a pyracylene patch, SWR2 occurs in a 15-atom patch with three hexagons. To date, only those two SWR subtypes have been observed experimentally.

Figure 3. Topology of skeletal transformations in fullerenes. The key C−C bonds (removed or rotated) are marked with ovals. The experimentally observed Stone−Wales rearrangements (SWRs) and C2 losses (C2Ls) are indicated with green arrows.

Chronologically, chlorination-promoted SWR transformation was first reported for D2-C76 where the likely starting IPR chlorides were characterized as well.13,46 The seven-step SWR1 sequence that produces the flattened non-IPR 18,917C76Cl24 with five FPPs and two vinylcoronene substructures (Figure 4, left) remains the longest pathway observed so far. More recently, D2-C76 was found to experience several other SWR pathways as well. A four-step SWR1 sequence yields 18,387 C76Cl30 with four FPPs (Figure 4, center) that can form concurrently with 18,917C76Cl24, despite that 18,387C76Cl30 is less

Figure 2. SWR transformations in the buckminsterfullerene C60. (a) Topological connections between different C60 cages in the isolated products. (b) Schlegel diagrams of the molecules with the addends shown (the CF3 groups, Cl and F atoms are indicated with black triangles, black and green circles, respectively). (c) Top and side projections of selected molecules showing adjacent pentagons (red) and a heptagon (blue).

captured in the form of 1809C60(CF3)10/14 after the follow-up trifluoromethylation.18 One more compound recovered by trifluoromethylation is a nonclassical C60(NC1)(CF3)15F with one heptagon, two fused pentagon pairs (FPPs) and one triple of sequentially fused pentagons (TSFP). It forms from 1809C60 via one further SWR2 step. The rest of the compounds were obtained as chlorides, including the highly similar 1810C60Cl24 and 1805C60Cl24 (both with four FPPs) and 1794C60Cl20 with five FPPs. They relate to 1809 C60 via one, two, and four SWR1 steps, respectively. Importantly, the pathway toward 1794C60Cl20 and its chlorination pattern are different from those of the other two chlorides, indicating the definitive role of chlorination patterns in regiochemistry of the transformation pathways and stabilization of the particular non-IPR fullerene cages.

Figure 4. Schlegel diagram and three projections of C78(NC2)Cl24. 1785

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research energetically advantageous because of lower number of chlorinated FPP sites and less favorable conjugated substructures.47 This example demonstrates that the regiochemistry of transformations is largely governed by the local kinetic factors which do not always correlate with the thermodynamic effects. DFT modeling of the SWR pathways in C76 has predicted several more possible transformation products and demonstrated that the regiochemical factors are largely connected with the rearranging chlorination patterns.47 However, the most recent study produced a yet different C76 cage missed by the modeling, C76(NC2)(CF3)14 (Figure 4, right), which features two heptagons, three FPPs, and two TSFPs.41 The C76(NC2) cage is related to the starting IPR-C76 via three SWR1 + two SWR2 steps and can give 18,917C76Cl24 via two further steps. Although DFT calculations predicted that 18,917 C76Cl24 is far more thermodynamically advantageous than the chlorides of C76(NC2), the experimental isolation of C76(NC2) provides another evidence of the role of the kinetic factors mediated by the chlorination patterns. The other reported cases of SWR transformations are relatively simpler. C82(3) was found to afford non-IPR 39,173 C82Cl28 with one FPP via two SWR1 steps (Figure 5).

Figure 6. Schlegel diagrams and projections of the non-IPR and nonclassical derivatives of C76.

IPR chloride C78Cl30,48 it was the much increased reaction time that eventually effected the transformation. 3.3. Cage Shrinkage via C2L Processes

Elimination of the C2 fragments from fullerene cages in conjunction with the formation of heptagons has already been studied by theorists.49 Both phenomena are important in the context of fullerene annealing in the arc-discharge plasma.50 The first observed chlorination-induced C2L process was a C2L1 elimination (see classification in Figure 3) from C86(16) upon chlorination with VCl4.14 The resulting C84(NC1)Cl32 formed alongside with a chloride of the parent IPR fullerene, C86(16)Cl28, which has a highly similar chlorination pattern and thus is regarded as a likely precursor of the C84(NC1)Cl32. Although the elimination mechanism in that particular case is not fully understood, it is clear that the C2L is favored by formation of a more planar 18-electron aromatic assembly with a heptagon. The most common C2L subtype is C2L3 for which the heptagon becomes flanked with two FPPs (Figure 3). Up to now, C2L3 eliminations have been observed in the same C86(16)21 and in C88(33),23 C90(28),24 C92(50),51 and C96(114).25 In all those cases, the eliminated fragment is chlorinated. In Figure 7, we present the starting compound and

Figure 5. Skeletal transformation of the IPR C82(3) (39,712C82Cl28) into the non-IPR 39,173C82Cl28 via two possible two-step SWR1 sequences. The relative energy values for the chlorinated molecules (before the parentheses) and for the pristine carbon cages (in parentheses) are given for all structures. The pathway via 37,349C82Cl28 is preferred due to higher exothermicity of the first stage.

Figure 7. Schlegel diagram presentation of a transformation from IPR C90(28)Cl24 to C88(NC1)Cl22 and C88(NC1)Cl24 via a C2L3 process. The leaving C2 unit is indicated with a small oval.

the products for the better studied case of C90(28), for which both the starting C90(28)Cl24 and the final C88(NC1)Cl22 and C88(NC1)Cl24 have been experimentally characterized.24 Moreover, a C2L2 elimination was assumed to occur in C96(80). The topological analysis revealed that the shortest pathway to the nonclassical C92(NC2)Cl32 product consists of two successive C2L2 steps.25 For the non-IPR fullerene cages, other kinds of C2L processes can occur. However, they are normally combined with the SWR rotations within branched multistep pathways, as discussed in the next section.

The transformations of C82 exemplify the criteria of choosing the more likely SWR sequence. The two SWR steps are formally permutable, but the calculated stability of the respective intermediate products is highly different, and it is assumed that the first SWR1 step is the more exothermic one that gives the 37,349C82Cl28 intermediate. Another example of SWR transformations was discovered for the isomeric mixture of C78(1−3). Prolonged chlorination with SbCl5 afforded flattened nonclassical C78(NC2)Cl24 with two heptagons and six FPPs (Figure 6).20 Interestingly, the chlorine atoms form a continuous closed loop that separates two large conjugated fragments from each other. According to the topological analysis, the most likely precursor of C78(NC2)Cl24 is C78(2), and the shortest transformation pathway includes four SWR1 and two SWR2 steps. Since previous chlorination experiments with C78(2) gave only the

3.4. Combined SWR and C2L Transformations

In Figure 8, we demonstrate one of the first examples of a branched transformation pathway that includes both SWR and C2L processes.28 The starting C100(18) was initially not known, and no intermediate products were isolated, but the 1786

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research

Figure 8. Shortest three-step pathways from a hypothetical IPR C100(18)Cl24 to the experimentally characterized C94(NC1)Cl22 and C96(NC3)Cl20. The fragments to be rotated or eliminated in the nearest step are encircled with ovals. Heptagons are marked as hp1− hp3.

Figure 9. Scheme of skeletal transformations of C88(33). The fragments to be rotated or eliminated in the nearest step are encircled with ovals. Heptagons are marked as hp1−hp3.

topological retroanalysis from the final products, C94(NC1)Cl22 and C96(NC3)Cl20, revealed C100(18) as the IPR isomer most closely related to both nonclassical chlorides. Later on, C100(18) was confirmed directly via isolation of C100(18)Cl28/30,26 demonstrating the relevance of the empirical principle of the shortest possible SWR/C2L pathways. The reconstructed pathway starts with two successive C2L3 steps that give C96(NC2) with two heptagons, two FPPs, and two TSFPs. Thereafter, the pathway branches. C2L5 elimination of one of the pentagon−pentagon junctions destroys one of the heptagons giving C94(NC1). Alternatively, an SWR2 rotation leads to C96(NC3) with a third heptagon, three FPPs, one TSFP, and a triple of directly fused pentagons (TDFP). Coformation of the two products suggested comparable activation barriers as confirmed by DFT. In the case of C88(33) (Figure 9), all the intermediate and terminal transformation products have been isolated and characterized.22 The initial C2L3 step affords C86(NC1)Cl26 intermediate, which can be followed by either another C2L3 elimination generating C84(NC2)Cl26) or an SWR1 yielding C86(NC1)Cl24. The latter compound can undergo an additional C2L2 elimination to afford two C84(NC2)Cl26 chlorides with the same carbon cage but slightly different chlorination patterns. Interestingly, under similar chlorination conditions, the other two abundant IPR isomers of C88, C88(7), and C88(17), only yield the corresponding chlorinated compounds without any skeletal transformations.52 C84(NC2)Cl30 synthesized via chlorination of a C92+C88 fraction with VCl4 features a unique carbon cage with two quadruples of sequentially fused pentagons (QSFPs) (Figure 10).23 The topological analysis identified a new isomer of C88, C88(3), as the most likely precursor that gives C84(NC2)Cl30 via two C2L2 and two SWR1 steps. According to the

Figure 10. Schlegel diagram and two projections of the C2C84(NC2)Cl30., which contains the first reported example of quadruples of sequentially fused pentagons (QSFP).

calculated relative energy and average C−Cl bond energy of the likely intermediates, the C2L steps precede the SWR ones. The skeletal transformations of C102 show a particularly diverse set of C2L subtypes. Initially, chlorination of C102 with VCl4/SbCl5 for several weeks produced only an SWR product, a non-IPR chloride 283,794C102Cl20 with two FPPs (Figure 11).29 Its topological analysis revealed the previously unknown IPR isomer C102(19) as the likely precursor, which transformed to non-IPR 283,794C102 cage via two SWR1 steps. The order of the steps predicted from the calculated exothermicity of the first step was confirmed later when the cocrystals of 283,794 C102Cl20 with the expected more stable intermediate, non-IPR 258,508C102Cl20, were obtained.30 Furthermore, under prolonged chlorination conditions, two further C2L products were captured, including a nonclassical C98(NC2)Cl26 and a non-IPR 185,115C96Cl28.45 C98(NC2)Cl26 can be obtained from C102(19) via successive C2L3 and C2L1 eliminations, while the formation of 185,115C96Cl28 requires C2L3, C2L4, and C2L5 sequential steps. Up to date, C102 represents the largest cage for which skeletal transformations were reliably proven experimentally. 1787

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research

effect the energetics and regiochemistry of the possible transformations. While the above general principles are obvious, detailed mechanistic understanding of the particular cases of transformations is often complicated by the existence of numerous alternatives and often lack of structural data for the starting compound and the intermediates. The skeletal processes are preceded by much faster chlorination of the starting cage18,20,46 and rapid rearrangements of the chlorination patterns via chlorine exchange with the chlorinating medium (i.e., chlorine dance15), imposing a possibility that the skeletal transformations start in some less abundant IPR chloride that offers lower activation barriers. In the multistep SWR/C2L processes, the chlorine dance and/or partial dechlorination or additional chlorination can further occur between the skeletal steps. A comprehensive computational survey of the whole range of possible chlorination patterns on the intermediate stages is unfeasible. While some low-cost empirical models can deliver successful predictions, verification at the DFT level will still be necessary, and it will have to encompass too many possible structures with different number of chlorine atoms. Hence, prediction of the likely chlorination patterns is inevitably based on the limited computational surveys that can be guided by some obvious empiric rules such as breakdown of the fullerene π-system into favorable aromatic fragments delimited by chlorine addends. While some of the thermodynamically advantageous chlorination patterns may thus get overlooked, this provides, at least, a helpful criterion to determine whether the given pattern can survive under the reaction conditions. In the known IPR chlorides that form under chlorination with SbCl5 or VCl4, such as C60Cl3016 and C70Cl28,17 the computed C−Cl binding energy falls in a rather narrow range. Hence, comparable or better average C−Cl binding energy in a hypothetical non-IPR or nonclassical chloride in question ensures stability of its chlorination pattern.

Figure 11. Schlegel diagram and projections of three molecules formed via skeletal transformations of C102(19).

4. CHLORINATION PATTERNS AND CHARACTERISTIC GEOMETRIC FEATURES According to DFT studies of the aforementioned compounds, the SWR processes in the chlorinated fullerenes are generally thermodynamically favored due to formation of the chlorinated pentagon−pentagon junctions. Typically, both sites of the FPP junctions are chlorinated, though in some cases one of the sites remains free because of being a member of some aromatic ring assembly. In the CF3 derivatives, the bulkier CF3 addends still tend to occupy both sites of the FPP junctions. In the triples and quadruples of sequentially fused pentagons, however, some of the sites at the pentagon−pentagon junctions always remain free. Otherwise, some of the pentagons would host as many as four Cl addends, which would be too sterically unfavorable. The same trends were previously found in chlorides of smaller non-IPR fullerenes.11 Unlike the TSFP patches, the TDFP ones allow for chlorination of all four sites of the pentagon− pentagon junctions, as seen in Figures 9 and 11. In general, the chlorine addends tend to arrange in such a way that they delimit various benzenoid and/or fused polycyclic conjugated substructures on the carbon cage, similarly to many IPR fullerene chlorides.34−38,40,41 However, the number and types of those substructures are always different and depend on the specific cage. Whenever the addends delimit the isolated double bonds or benzenoid rings, we observe the respective characteristic C−C bond lengths of 1.32−1.34 or 1.39−1.40 Å. The stronger C−Cl bonds at the pentagon−pentagon junctions are, accordingly, noticeably shorter than the other C−Cl bonds. For example, the 10 C−Cl bonds at the FPP junctions in the D5-1794C60Cl20 have the average length of 1.776 vs 1.80−1.83 Å C−Cl bonds typically observed at the pentagon−hexagon−hexagon junctions.33−40

5.2. DFT Modeling of the SWR and C2L Mechanisms

Earlier DFT modeling of an SWR act in C60 identified the two possible mechanisms: the thermally forbidden concerted pathway where the rotated C2 fragment transiently becomes a triple bond and the asymmetric, carbene, pathway.4 In the latter case, one site of the rotated fragment becomes a carbene center while the other becomes an sp3 one. Both options were found to have extremely high activation barriers of ca. 7 eV. Later, it was shown that the activation barrier of the asymmetric pathway can be reduced to ca. 5 eV by stabilizing the carbene center with some radical.53 It turns out that the migrating chlorine addends in chlorinated fullerenes provide far better stabilization of the asymmetric transition state. The mechanism of asymmetric SWR rotation of a chlorinated C2 fragment in the presence of SbCl5 is shown in Figure 12 for the SWR2 case. In order to maintain advantageous valence-saturated electronic structure across the reaction path, chlorine is transferred between the two sites of the rotated fragment by SbCl5 so that the intermediate structure features a CCl2 bridge. The transition states appear as closed-shell ion pairs with stable SbCl6− anions. In the first example of chlorination-promoted SWRs in D2C76, the barrier of the limiting step was calculated to be 240 kJ mol−1, in agreement with the experimental conditions,13 while the combined exothermic effect of the seven-step transformation into 18,917C76Cl24 exceeded 400 kJ mol−1.46 Later, it

5. QUANTUM-CHEMICAL MODELING 5.1. General Remarks

The sources of stabilization of the non-IPR cages are (i) formation of the chlorinated pentagon-pentagon junctions with increased chlorination energy due to their lower stability in the sp2-hybridized state and, sometimes, (ii) formation of more favorable aromatic substructures than those present in the precursory IPR cage. Depending on the topology of the fullerene cage, the chlorination patterns, in their turn, crucially 1788

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research

portions of fullerene cages as sketched in Figure 13.24 Its first stage is similar to formation of an SWR intermediate, but the

Figure 12. Two-substep Stone−Wales rearrangement (SWR2 subtype; see Figure 3) in chlorinated fullerenes in the presence of the chlorinating agent. Acronym “TS” denotes the ion-pair transition states.

was found that the barriers are often considerably lower, even below 200 kJ mol−1, evidencing that chlorine is a particularly advantageous stabilizing moiety.28 Indeed, its moderate binding energy with fullerenes provides lability at elevated temperatures without decomposition, and ready formation of stable SbCl6− anions enables efficient chlorine migration. The C60 case that required higher reaction temperatures18 clearly illustrates the role of the chlorination patterns. High stability of the likely precursor, IPR D3d-C60Cl30,16 makes the first SWR1 step highly endothermic, unless the chlorination pattern rearranges in such a way so as to stabilize the resulting 1809 C60. It is also of high activation energy of 291 kJ mol−1. However, with a hypothetic closed-loop chlorination pattern like in the final 1794C60Cl20 (see Figure 2), the activation energy would drop to more common 227 kJ mol−1, and the overall exothermic effect of the five-step rearrangement would reach as much as 599 kJ mol−1. Thus, the buckminsterfullerene in itself is not particularly stable against the skeletal transformations, and the higher reaction temperatures are possibly required to increase the abundance of those slightly less stable chlorination patterns that favor the transformations. Two further comments on the chlorination-promoted SWR mechanism are to be made. First, chloride transfer between the starting or final fullerene cage and SbCl5/SbCl6− can formally constitute a separate reaction substep. However, its activation barrier is typically within only 1 eV, indicatively of the ease of the SbCl5-mediated chlorine migration. Second, only one of the sites of the rotated C2 fragment should necessarily carry a chlorine addend: the one to become a CCl2 bridge. The other chlorine atom can, in principle, be brought by SbCl6− from/to some other position. However, according to our observations, it is usually advantageous that both sites are chlorinated and, moreover, are within a longer continuous chain of chlorinated sites, as shown in Figure 12. The C2L processes are generally harder to rationalize. In some cases, they involve the C2 fragments that are likely not chlorinated, as it is, for example, in the case of C86(16), and the nature of the catalytic effect of the chlorinating medium remains obscure. Moreover, it is not understood why the C2L processes are particularly favored by VCl4. Our DFT modeling revealed no fullerene−VCl4/VCl3 interactions other than normal chlorination/dechlorination and chloride transfer through transiently forming VCl5−. However, we were able to propose a general C2L mechanism for the chlorinated

Figure 13. General branched pathway of C2 elimination. The structures with black numerals represent the minima on the potential energy surface, and the transition states between them are denoted as “TS#”. The forming and cleaving bonds at the transition states are shown with red dashes.

atom that develops four C−C bonds does not belong to the C2 fragment to be abstracted. Then, the pathway can branch, giving either a doubly bonded CCl2 moiety or a −CCl2− CCl2-bridge. The bridge is subsequently eliminated, either in a concerted manner or asymmetrically, via the −C2Cl3 and −C2Cl5 states, and the CCl2 branch of the pathway subsequently merges into the asymmetric route. The calculations for the VCl4-promoted C90-to-C88 transformation produced the limiting activation energy of 230−250 kJ mol−1 with respect to the starting structure.24 In a later work,28 elimination of a pentagon−pentagon junction from C96(NC2)Cl22 in the presence of SbCl5 to give C94(NC1)Cl22 was calculated to have far lower barrier of 162 kJ mol−1 along the concerted pathway. As with the SWRs, the computational estimates for the C2L processes agree well with the experimental conditions.

6. CONCLUSION AND OUTLOOK On the basis of the survey of the chlorination-promoted skeletal transformations of fullerenes in the range of C60−C102 presented above, one can see that such phenomena become a new forefront of the fullerene chemistry. In particular, our recent discovery of the transformations in the abundant C60 signifies the potential applications of non-IPR fullerenes in addition to their fundamental importance. However, a thorough and in-depth understanding of the skeletal transformations of fullerenes still requires further studies to cover the entire fullerene family especially the second most abundant member C70. Even for the already studied systems, there may 1789

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research be overlooked products, and variations of the reaction time and temperature may bring about further discoveries. In view of that, the following two research directions are particularly worthwhile to pursue in the future:

many articles on synthesis and structural studies of inorganic and organometallic compounds by X-ray and neutron diffraction. His current research interests are focused on the structural chemistry of the derivatives of fullerenes including the higher ones.

■

6.1. Recovery of the Pristine Non-IPR and Nonclassical Fullerene Isomers

ACKNOWLEDGMENTS The authors acknowledge financial supports from the National Key Research and Development Program of China (2017YFA0402800), National Natural Science Foundation of China (51572254) [to S.Y.], and the Russian Foundation for Basic Research (Grant 19-03-00733) [to S.I.T.]. This work was carried out using the equipment of the shared research facilities of HPC computing resources at Lomonosov Moscow State University.

The non-IPR fullerene cages are expected to possess unusual electronic and chemical properties. The fused pentagon assemblies will give rise to narrower HOMO−LUMO gap and related enhancement in both donor and acceptor properties. In addition, higher reactivity of the junctions will lead to regioselectivity in the exohedral functionalization. The first attempts to recover a pristine non-IPR 1809C60 cage from its chloride 1809C60Cl8 obtained via a dopant-modified arcdischarge synthesis were of limited success.54 The recently discovered transformation of the readily available C60 will enhance availability of its non-IPR and nonclassical derivatives18 for further experimentation. Their lability will remain a fundamental obstacle, but circumventable by means of direct substitution of addends without intermediate recovery of pristine fullerene cages, similar to substitution in the non-IPR 1809 C60Cl8.9 The recovery of some of the transformed cages via in situ trifluoromethylation of the chlorinated products reported recently18,41 is a ready example of such approach.

■

(1) Schmalz, T. G.; Seitz, W. A.; Klein, D. J.; Hite, G. E. Elemental carbon cages. J. Am. Chem. Soc. 1988, 110, 1113−1127. (2) Cross, R. J.; Saunders, M. Transmutation of fullerenes. J. Am. Chem. Soc. 2005, 127, 3044−3047. (3) Stone, A. J.; Wales, D. J. Theoretical studies of icosahedral C60 and some related species. Chem. Phys. Lett. 1986, 128, 501−503. (4) Bettinger, H. F.; Yakobson, B. I.; Scuseria, G. E. Scratching the surface of buckminsterfullerene: The barriers for Stone-Wales transformation through symmetric and asymmetric transition states. J. Am. Chem. Soc. 2003, 125, 5572−5580. (5) Hirsch, A.; Brettreich, M. Fullerenes. Chemistry and Reactions; Wiley-VCH: Weinheim, 2005. (6) Zettergren, H.; Alcamí, M.; Martín, F. Stable non-IPR C60 and C70 fullerenes containing a uniform distribution of pyrenes and adjacent pentagons. ChemPhysChem 2008, 9, 861−866. (7) Xie, S.-Y.; Gao, F.; Lu, X.; Huang, R.-B.; Wang, C.-R.; Zhang, X.; Liu, M.-L.; Deng, S.-L.; Zheng, L.-S. Capturing the labile fullerene[50] as C50Cl10. Science 2004, 304, 699. (8) Wang, C. R.; Shi, Z. Q.; Wan, L. J.; Lu, X.; Dunsch, L.; Shu, C. Y.; Tang, Y. L.; Shinohara, H. C64H4: Production, isolation, and structural characterizations of a stable unconventional fulleride. J. Am. Chem. Soc. 2006, 128, 6605−6610. (9) Tan, Y.-Z.; Liao, Z.-J.; Qian, Z.-Z.; Chen, R.-T.; Wu, X.; Liang, H.; Han, X.; Zhu, F.; Zhou, S.-J.; Zheng, Z.; Lu, X.; Xie, S.-Y.; Huang, R.-B.; Zheng, L.-S. Two Ih-symmetry-breaking C60 isomers stabilized by chlorination. Nat. Mater. 2008, 7, 790−794. (10) Tan, Y.-Z.; Xie, S.-Y.; Huang, R.-B.; Zheng, L.-S. The stabilization of fused pentagon fullerene molecules. Nat. Chem. 2009, 1, 450−460. (11) Tan, Y.-Z.; Li, J.; Zhu, F.; Han, X.; Jiang, W.-S.; Huang, R.-B.; Zheng, Z.; Qian, Z.-Z.; Chen, R.-T.; Liao, Z.-J.; Xie, S.-Y.; Lu, X.; Zheng, L.-S. Chlorofullerenes featuring triple sequentially fused pentagons. Nat. Chem. 2010, 2, 269−273. (12) Troshin, P. A.; Avent, A. G.; Darwish, A. D.; Martsinovich, N.; Abdul-Sada, A. K.; Street, J. M.; Taylor, R. Isolation of two sevenmembered ring C58 fullerene derivatives: C58F17CF3 and C58F18. Science 2005, 309, 278−281. (13) Ioffe, I. N.; Goryunkov, A. A.; Tamm, N. B.; Sidorov, L. N.; Kemnitz, E.; Troyanov, S. I. Fusing pentagons in a fullerene cage via chlorination: IPR D2-C76 rearranges into non-IPR C76Cl24. Angew. Chem., Int. Ed. 2009, 48, 5904−5907. (14) Ioffe, I. N.; Chen, C.; Yang, S. F.; Sidorov, L. N.; Kemnitz, E.; Troyanov, S. I. Chlorination of C86 to C84Cl32 with nonclassical heptagon-containing fullerene cage formed by cage shrinkage. Angew. Chem., Int. Ed. 2010, 49, 4784−4787. (15) Fowler, P. W.; Manolopoulos, D. E. An Atlas of Fullerenes; Clarendon: Oxford, U.K., 1995. (16) Troshin, P. A.; Lyubovskaya, R. N.; Ioffe, I. N.; Shustova, N. B.; Kemnitz, E.; Troyanov, S. I. Synthesis and structure of the highly chlorinated [60]fullerene C60Cl30 with drum-shaped carbon cage. Angew. Chem., Int. Ed. 2005, 44, 234−237.

6.2. Development of New Synthetic Protocols

Up to now, the skeletal transformation experiments utilized the same basic composition of the reaction mixture: a fullerene and an excess of a chlorinating agent. By using more complex compositions, one can expect deeper skeletal transformations or products with larger distributions in the fullerene cage without a risk of its further degradation. On the other hand, it would be also interesting to investigate the possibility of skeletal transformations by the fluorination under milder conditions than that used in the C58F18 case.12 While the C−F bonds are generally stronger than the C−Cl ones, fluorine can also form highly stable complex anions like SbF6− to mediate the necessary fluorine migration.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Shangfeng Yang: 0000-0002-6931-9613 Sergey I. Troyanov: 0000-0003-1663-0341 Notes

The authors declare no competing financial interest. Biographies Shangfeng Yang is currently a Full Professor of University of Science and Technology of China sponsored by the “Hundreds of Talents Programme” of Chinese Academy of Sciences. He obtained his Ph.D. in 2003 and then joined IFW-Dresden, Germany as a Humboldt Fellow. His present interests include the syntheses of fullerene-based nanocarbons toward applications in energy conversion and storage. Ilya N. Ioffe obtained his Ph.D. in 2002 and habilitated in 2012 at the Moscow State University. His present interests are focused on the theoretical chemistry of fullerenes and computational photochemistry. Sergey I. Troyanov obtained his Ph.D. in 1973 and habilitated in 1991 at the Moscow State University. Prof. Troyanov is the author of 1790

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research

connectivities and chlorination patterns. Chem. - Eur. J. 2011, 17, 10662−10669. (35) Tamm, N. B.; Troyanov, S. I. New isolated-pentagon-rule isomer of C92 isolated as trifluoromethyl and chlorido derivatives: C92(38)(CF3)14/16 and C92(38)Cl20/22. Inorg. Chem. 2015, 54, 10527− 10529. (36) Tamm, N. B.; Yang, S. F.; Wei, T.; Troyanov, S. I. Five isolated Pentagon Rule Isomers of higher fullerene C94 captured as chlorides and CF3 derivatives: C94(34)Cl14, C94(61)Cl20, C94(133)Cl22, C94(42)(CF3)16, and C94(43)(CF3)18. Inorg. Chem. 2015, 54, 2494− 2496. (37) Yang, S. F.; Wei, T.; Kemnitz, E.; Troyanov, S. I. Four isomers of C96 fullerene structurally proven as C96Cl22 and C96Cl24. Angew. Chem., Int. Ed. 2012, 51, 8239−8242. (38) Jin, F.; Yang, S. F.; Troyanov, S. I. New isolated-pentagon-rule isomers of fullerene C98 captured as chloro derivatives. Inorg. Chem. 2017, 56, 4780−4783. (39) Fritz, M. A.; Kemnitz, E.; Troyanov, S. I. Capturing an unstable C100 fullerene as chloride, C100(1)Cl12, with a nanotubular carbon cage. Chem. Commun. 2014, 50, 14577−14580. (40) Yang, S. F.; Wang, S.; Troyanov, S. I. The most stable isomers of giant fullerenes C102 and C104 captured as chlorides, C102(603)Cl18/20 and C104(234)Cl16/18/20/22. Chem. - Eur. J. 2014, 20, 6875− 6878. (41) Tamm, N. B.; Brotsman, V. A.; Markov, V. Yu.; Kemnitz, E.; Troyanov, S. I. Chlorination-promoted transformation of IPR C76 discovered via trifluoromethylation under formation of non-IPR C76(CF3)nFm. Dalton Trans 2018, 47, 6898−6902. (42) Boltalina, O. V.; Popov, A. A.; Kuvychko, I. V.; Shustova, N. B.; Strauss, S. H. Perfluoroalkylfullerenes. Chem. Rev. 2015, 115, 1051− 1105. (43) Laikov, D. N. Fast evaluation of density functional exchangecorrelation terms using the expansion of the electron density in auxiliary basis sets. Chem. Phys. Lett. 1997, 281, 151−156. (44) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (45) Granovsky, A. A. Firefly, version 8.2; http://classic.chem.msu. su/gran/firefly/index.html. (46) Ioffe, I. N.; Mazaleva, O. N.; Chen, C.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. C76 fullerene chlorides and cage transformations. Structural and theoretical study. Dalton Trans 2011, 40, 11005− 11011. (47) Sudarkova, S. M.; Mazaleva, O. N.; Konoplev-Esgenburg, R. A.; Troyanov, S. I.; Ioffe, I. N. Versatility of chlorination-promoted skeletal transformation pathways in the C76 fullerene. Dalton Trans. 2018, 47, 4554−4559. (48) Kemnitz, E.; Troyanov, S. I. Chlorides of isomeric C78 fullerenes: C78(1)Cl30, C78(2)Cl30, and C78(2)Cl18. Mendeleev Commun. 2010, 20, 74−76. (49) Murry, R. L.; Strout, D. L.; Odom, G. K.; Scuseria, G. E. Role of sp3 carbon and seven-membered rings in fullerene annealing and fragmentation. Nature 1993, 366, 665−667. (50) Zhang, J.; Bowles, F. L.; Bearden, D. W.; Ray, W. K.; Fuhrer, T.; Ye, Y.; Dixon, C.; Harich, K.; Helm, R. F.; Olmstead, M. M.; Balch, A. L.; Dorn, H. C. A missing link in the transformation from asymmetric to symmetric metallofullerene cages implies a top-down fullerene formation mechanism. Nat. Chem. 2013, 5, 880−885. (51) Guan, R.; Jin, F.; Yang, S. F.; Tamm, N. B.; Troyanov, S. I. Stable C92(26) and C92(38) as well as unstable C92(50) and C92(23) IPR isomers revealed by chlorination of C92 fullerene. Inorg. Chem. 2019, 58, 5393−5396. (52) Wang, S.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. Unusual chlorination patterns of three IPR isomers of C88 fullerene in C88(7)Cl12/24, C88(17)Cl22, and C88(33)Cl12/14. Chem. - Asian J. 2016, 11, 77−80. (53) Alder, R. W.; Harvey, J. N. Radical-promoted Stone-Wales rearrangements. J. Am. Chem. Soc. 2004, 126, 2490−2494. (54) Chen, R.-T.; Zhou, S.-J.; Liang, H.; Qian, Z.-Z.; Li, J.-M.; He, Q.; Zhang, L.; Tan, Y.-Z.; Han, X.; Liao, Z.-J.; Weng, W.-Z.; Xie, S.-Y.;

(17) Troyanov, S. I.; Shustova, N. B.; Ioffe, I. N.; Turnbull, A. P.; Kemnitz, E. Synthesis and structural characterization of highly chlorinated C70, C70Cl28. Chem. Commun. 2005, 72−74. (18) Brotsman, V. A.; Tamm, N. B.; Markov, V. Yu.; Ioffe, I. N.; Goryunkov, A. A. S.; Kemnitz, E.; Troyanov, S. I. Rebuilding C60: chlorination-promoted transformations of the buckminsterfullerene into pentagon-fused C60 derivatives. Inorg. Chem. 2018, 57, 8325− 8331. (19) Ioffe, I. N.; Mazaleva, O. N.; Sidorov, L. N.; Yang, S. F.; Wei, T.; Kemnitz, E.; Troyanov, S. I. Skeletal transformation of Isolated Pentagon Rule (IPR) fullerene C82 into non-IPR C82Cl28 with notably low activation barriers. Inorg. Chem. 2012, 51, 11226−11228. (20) Brotsman, V. A.; Ignat’eva, D. V.; Troyanov, S. I. Chlorinationpromoted transformation of isolated pentagon rule C78 into fusedpentagons-and heptagons-containing fullerenes. Chem. - Asian J. 2017, 12, 2379−2382. (21) Wei, T.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. Two successive C2 losses from C86 fullerene upon chlorination with the formation of non-classical C84Cl30 and C82Cl30. Chem. - Asian J. 2015, 10, 559−562. (22) Yang, S. F.; Wei, T.; Scheurell, K.; Kemnitz, E.; Troyanov, S. I. Chlorination-promoted skeletal cage transformations of C88 fullerene by C2 losses and a C-C bond rotation. Chem. - Eur. J. 2015, 21, 15138−15141. (23) Jin, F.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. Skeletal transformation of a classical fullerene C88 into a nonclassical fullerene chloride C84Cl30 bearing quaternary sequentially fused pentagons. J. Am. Chem. Soc. 2017, 139, 4651−4654. (24) Ioffe, I. N.; Mazaleva, O. N.; Sidorov, L. N.; Yang, S. F.; Wei, T.; Kemnitz, E.; Troyanov, S. I. Cage shrinkage of fullerene via a C2 loss: from IPR C90(28)Cl24 to non-classical, heptagon-containing C88Cl22/24. Inorg. Chem. 2013, 52, 13821−13823. (25) Yang, S. F.; Wei, T.; Wang, S.; Ioffe, I. N.; Kemnitz, E.; Troyanov, S. I. Structures of chlorinated fullerenes, IPR C96Cl20 and non-classical C94Cl28 and C92Cl32: Evidence of the existence of three new isomers of C96. Chem. - Asian J. 2014, 9, 3102−3105. (26) Wang, S.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. New isolated-pentagon-rule and skeletally transformed isomers of C100 fullerene identified by structure elucidation of their chloro derivatives. Angew. Chem., Int. Ed. 2016, 55, 3451−3454. (27) Yang, S. F.; Wang, S.; Kemnitz, E.; Troyanov, S. I. Chlorination of IPR C100 fullerene affords unconventional C96Cl20 with a nonclassical cage containing three heptagons. Angew. Chem., Int. Ed. 2014, 53, 2460−2463. (28) Ioffe, I. N.; Yang, S. F.; Wang, S.; Kemnitz, E.; Sidorov, L. N.; Troyanov, S. I. C100 is converted into C94Cl22 via three chlorinationpromoted C2 losses under formation and elimination of cage heptagons. Chem. - Eur. J. 2015, 21, 4904−4907. (29) Yang, S. F.; Wei, T.; Wang, S.; Ignat’eva, D. V.; Kemnitz, E.; Troyanov, S. I. The first structural confirmation of a C102 fullerene as C102Cl20 containing a non-IPR carbon cage. Chem. Commun. 2013, 49, 7944−7946. (30) Mazaleva, O. N.; Ioffe, I. N.; Jin, F.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. Experimental and theoretical approach to variable chlorination-promoted skeletal transformations in fullerenes: The case of C102. Inorg. Chem. 2018, 57, 4222−4225. (31) Wang, S.; Yang, S. F.; Kemnitz, E.; Troyanov, S. I. New giant fullerenes identified as chloro derivatives: Isolated-pentagon-rule C108(1771)Cl 12 and C 106 (1155)Cl24 as well as non-classical C104Cl24. Inorg. Chem. 2016, 55, 5741−5743. (32) Yang, S. F.; Wei, T.; Troyanov, S. I. A New isomer of pristine higher fullerene Cs-C82(4) captured by chlorination as C82Cl20. Chem. - Asian J. 2013, 8, 351−353. (33) Lanskikh, M. A.; Tamm, N. B.; Sidorov, L. N.; Troyanov, S. I. Capturing C84 isomers as chlorides and pentafluoroethyl derivatives: C84Cl22 and C84(C2F5)12. Inorg. Chem. 2012, 51, 2719−2721. (34) Troyanov, S. I.; Yang, S. F.; Chen, C.; Kemnitz, E. Six IPR isomers of C90 fullerene captured as chlorides: Carbon cage 1791

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792

Article

Accounts of Chemical Research Huang, R.-B.; Zheng, L.-S. C2v-symmetric C60 isomer in the gas phase: Experimental evidence against buckminsterfullerene (Ih-C60). J. Phys. Chem. C 2009, 113, 16901−16905.

1792

DOI: 10.1021/acs.accounts.9b00175 Acc. Chem. Res. 2019, 52, 1783−1792