Research ArticleMATERIALS SCIENCE

Rational synthesis of an atomically precise carboncone under mild conditions

See allHide authors and affiliations

Science Advances  23 Aug 2019:
Vol. 5, no. 8, eaaw0982
DOI: 10.1126/sciadv.aaw0982

Abstract

Carboncones, a special family of all-carbon allotropes, are predicted to have unique properties that distinguish them from fullerenes, carbon nanotubes, and graphenes. Owing to the absence of methods to synthesize atomically well-defined carboncones, however, experimental insight into the nature of pure carboncones has been inaccessible. Herein, we describe a facile synthesis of an atomically well-defined carboncone[1,2] (C70H20) and its soluble penta-mesityl derivative. Identified by x-ray crystallography, the carbon skeleton is a carboncone with the largest possible apex angle. Much of the structural strain is overcome in the final step of converting the bowl-shaped precursor into the rigid carboncone under mild reaction conditions. This work provides a research opportunity for investigations of atomically precise single-layered carboncones having even higher cone walls and/or smaller apex angles.

INTRODUCTION

As new structural forms of carbon, carboncones were first observed as structural defects on the endcaps of deformed carbon nanotubes (CNTs) by Iijima and co-workers in 1992 (1). A series of carboncones with one, two, three, four, or five pentagons on the cone cap were subsequently observed by electron microscopy (Fig. 1A) (2, 3). Carboncones can be considered as intermediate species along the continuous formal transition (2) from fullerenes to graphenes through a tubular-like stage, resembling a zero-dimensional (0D) cluster at the apex, then proceeding to a 1D tube, and, finally, approaching a 2D graphene. Largely owing to their unique conical shapes and hollow structures, carboncones hold particular promise for applications such as scanning probe tips (4), electron field emitters (5), and gas storage materials (6). Some chemical-physical techniques, including chemical vapor deposition (4), laser ablation (7), arc discharge (1), and pyrolysis of organic compounds (3), have been developed to produce multilayered carboncones, but yields of those carboncones are rather low. Moreover, mixtures of carboncones with varying cone angles and heights were typically produced together with impurities of carbon discs and CNTs in carbon soots (3). It remains challenging to get rid of the impurities from these mixtures and to synthesize a single-layered carboncone. Research of carboncones has been severely hampered by the limitations of the previous synthetic technology in the past 20 years. Consequently, insight into the nature of well-defined single-layered carboncones has been inaccessible, and controllable methods for synthesizing structurally uniform carboncones are strongly desired.

Fig. 1 Structures of carboncone.

(A) Illustration of all five types of carboncone[n,m] (n = 1 to 5, m = 3 and 4). The bonds of pentagon are marked with green lines. The circles of fused rings are painted with orange, pink, yellow, and white from tips to rims, respectively. The black dot points to the cone tip. (B) The Oak Ridge thermal-ellipsoid plot (ORTEP) diagram of the penta-mesityl carboncone[1,2] derivative (1b) at 50% probability with hydrogen atoms and solvent molecules omitted for clarity (top view), and stick model diagram of crystallographic structure of 1b with five mesityl groups, hydrogen atoms, and solvent molecules omitted for clarity (side view). The cone height and the apex angle in the crystal are displayed.

Herein, we reported a bottom-up organic chemical approach to prepare a carboncone molecule 1a (C70H20) and its penta-mesityl derivative 1b under mild reaction conditions. The carboncone framework of the newly synthesized molecules with a single pentagonal cone cap and a hexagonal lattice cone wall has been unambiguously identified by x-ray crystallography (Fig. 1B).

To better communicate about carboncones with varying cone caps and cone walls, we proposed a nomenclature of carboncone[n,m]. Here, “n” refers to the number of pentagons on the cone cap (“n” determines the apex angle of the cone), and “m” refers to the number of intact circles consisting of fused pentagonal/hexagonal rings (“m” determines the height of the cone in some senses). Shown in Fig. 1A are a series of typical carboncone[n,m] in which n = 1 to 5 and m = 3 or 4. The newly synthesized carboncone skeleton can be defined as a carboncone[1,2], in which the “1” indicates one pentagon on the cone cap and the “2” reflects two intact circles of fused hexagonal rings between the cone tip and the cone rim (Fig. 1A). Note that the nomenclature of carboncone[n,m] does not exclude structures having irregular or incomplete rims. Strictly speaking, for example, the skeleton of carboncone 1 is a pentanaphtho-carboncone[1,2]. Similar inclusiveness is likewise found in the definition for CNT [n,m] (8). As outlined in Fig. 1, the tip of the carboncone can be a point in the center of a pentagonal/hexagonal ring or a point at the middle of the central carbon-carbon bond. According to this definition, the well-known corannulene molecule can be defined as a “cone cap,” including only a single circle of fused hexagonal rings around the cone tip, rather than as a carboncone, which must have at least one circle of cone wall in addition to a cone cap. Notably, the carboncone with five pentagons on the cone cap, such as carboncone[5,4], have no centrosymmetric point; as a result, the carboncone wall between the cone tip and the cone rim is uneven, and the carboncone[5,4] looks tilted.

RESULTS

Synthesis and structures of carboncone[1,2] derivatives 1

As shown in Fig. 2, the carboncone[1,2] derivatives 1a and 1b were synthesized by three-step organic chemical processes starting from corannulene 2. Direct borylation of 2 with B2(pin)2 afforded 1,3,5,7,9-pentakis(Bpin)corannulene 3 (9) as the major product. Then, fivefold Suzuki-Miyaura coupling of 3 with 1-bromonaphthalene 4a or 1-bromo-4-mesitylnaphthalene 4b gave sym-penta(1-naphthyl)corannulene 5a or sym-penta(4-mesityl-1-naphthyl)corannulene 5b in moderate yields (figs. S1 to S6). Last, when cyclodehydrogenation reaction conditions were applied to the bowl-shaped polycyclic aromatic hydrocarbons (PAHs) 5 (10), completely closed products 1 were obtained by stitching together the five appendages of 5.

Fig. 2 Three-step organic synthesis of the carboncone[1,2] derivatives 1 from corannulene 2.

Conditions: (i.) (Ir(OMe)cod)2 [20 mole percent (mol %)], B2(pin)2 (6.5 equiv.), 4,4′-dimetylpyridine (40 mol %), sodium methoxide (10 mol %), cyclohexane, 85°C, 4 days; (ii.) 1-bromonaphthalene or 1-bromo-4-mesitylnaphthalene (10 equiv.), Pd(PPh3)4 (25 mol %), Cs2CO3 (30 equiv.), toluene/H2O (2:1), 100°C, 24 hours; (iii.) DDQ (20 equiv.), TfOH/CH2Cl2 (1:20), 10 min. cod, 1,5-cyclooctadiene; pin, pinacolato; cat., catalyst; DDQ, 2,3-dichloro-5,6-dicyano-1,4-benzoquinone.

The hydrogen-terminated carboncone[1,2] 1a is rather poorly soluble in common organic solvents, likely owing to the strong van der Waals attractions associated with large-area intermolecular contacts between curved faces. Owing to such a low solubility, carboncone 1a was characterized only by high-resolution mass spectrometry with matrix-assisted laser desorption ionization (MALDI) (fig. S7), which confirmed the molecular formula of 1a as C70H20. In addition to the major component of the completely closed 1a, a mixture of some unidentified products likely arising from incomplete closure were also detected in the mass spectra (fig. S7). To improve the solubility of 1a, we introduced mesityl groups at the 4-postion of the naphthyl units. As anticipated, the corresponding ring closure product 1b was synthesized in moderate yield and exhibited much improved solubility compared with 1a.

With sufficient solubility in common organic solvents (e.g., dichloromethane, carbon disulfide, and toluene), the carboncone[1,2] derivative 1b can be characterized by routine techniques, such as nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (figs. S8 to S11). The 1H NMR spectrum of 1b reveals the presence of five different types of aromatic protons (three singlets and two doublets) (fig. S9), indicating that 1b has C5 symmetry. The calculated 1H NMR spectrum of carboncone 1b is in good agreement with the experimentally observed spectrum (fig. S10). As expected, in the aromatic region, the 13C NMR spectrum shows 20 signals: 138.80, 138.71, 137.84, 137.47, 137.28, 136.97, 130.05, 129.02, 128.93, 128.65, 127.98, 127.91, 125.18, 125.08, 124.73, 124.68, 124.57, 123.72, 123.66, and 123.53 ppm (fig. S11).

A single crystal of 1b suitable for x-ray analysis was obtained by slow diffusion of hexane into a carbon disulfide solution of 1b (table S1). In accordance with NMR spectra (figs. S9 to S11), the crystallographic structure of 1b confirms carboncone[1,2] having a C5 symmetrically conical shape with one pentagon embedded in 25 hexagons (Fig. 1B). The unique conical curvature of carboncone[1,2] distinguishes it from the dozens of known buckybowls (11) and fullerenes (12) that also exhibit positive geodesic curvature.

In the parent carboncone[1,2] of 1b, 70 sp2-carbon atoms are completely conjugated. The experimentally determined apex angle of 1b is 117.6° (Fig. 1B), which is slightly larger than that of the large cone shown in the topological model (112.9°) (1, 2). The central corannulene moiety adopts a cone-shaped geometry with a depth of 1.16 Å, which is substantially deeper than that of the parent corannulene (0.86 Å). The cone height, defined as the distance between the center of the pentagon and the mean plane of the 10 furthest rim carbons, is 3.29 Å (Fig. 1B). The p-orbital axis vector (POAV) angle of the carbon atoms in the cone tip of 1b (10.5°) is also larger than that in the parent corannulene (8.2°) (13) but smaller than that in the fullerene C60 (11.6°) (13). The steric hindrance provided by the five mesityl groups in 1b prevents the molecules from stacking in a concave-to-convex orientation (14). Rather than forming a columnar arrangement in the crystal, a staggered packing is adopted (fig. S12). In addition, the cone-to-cone inversion energy of 1a and 1b (166.4 and 167.0 kcal/mol, respectively) through an S-shaped transition state was estimated by density function theory (DFT) calculations, and is 188.8 kcal/mol for 1a if through a hypothetical flat transition state structure (see figs. S13 and S14, table S2, and Computational methods in Materials and Methods). These large inversion energies exceed the bond dissociation energies of even the strongest Csp2─Csp2 bonds and indicate that the geometric structure of carboncone[1,2] must be quite rigid.

Theoretical understanding of the ring closure reaction from the precursor 5 toward 1

Theoretical calculations were performed to understand the final ring closure reaction. The strain energy of 1b (98.4 kcal/mol) and the strain increase in the final closure step (43.5 kcal/mol) were estimated by DFT calculations (see fig. S15 and Computational methods in Materials and Methods).

To understand why the ring closure from the precursor 5 toward 1 can be completed under these mild reaction conditions, we proposed a rational Scholl ring closure reaction pathway leading to carboncone[1,2] (1a) with the lowest barriers in each step of cyclization (Fig. 3). For the most likely pathway shown in Fig. 3 (blue and red hexagons closing alternately but not always adjacent), the activation energies associated with the formation of the red hexagons appear to be lower than those for the formation of the blue hexagons. To probe the differences between the red and the blue hexagons, we calculated nucleus-independent chemical shifts (NICS) (15) for the new rings in each of the assumed intermediates (Fig. 3). The calculation results show that the NICS values of the red rings range from −6.28 to −9.94, indicating a strong aromatic character. The blue hexagons, on the other hand, are revealed to be nonaromatic, with NICS values ranging from +2.38 to +6.98. These differences in value of NICS suggest that formation of the aromatic red hexagons makes favorable contributions to reduce barriers of the steps of cyclization.

Fig. 3 A rational ring closure pathway from the precursor 5a toward 1a calculated by a dication pathway.

Here, ∆E represents the energy barrier for each step of cyclization. The NICS values of the rings in all intermediate products from A1 to A11 are presented.

Energy barriers for each step of cyclization were calculated by a dication pathway (16) in the polarized continuum model with dichloromethane solvent (figs. S16 to S23). All possible modes of six-membered ring formation in 10 total cyclizations were considered, and the corresponding energy barriers were obtained (shown in figs. S16 to S22). Taking the calculations of ring formation on A2 as an example, the formation of an adjacent aromatic hexagon (∆E2–1 = 11.6 kcal mol−1) is much easier than the formation of the other nonaromatic hexagons (from A2 to A2-2, A2-3, A2-4, or A2-5) (fig. S16). Note that the pathway of the first formation of all five nonaromatic blue hexagons (rather than the five aromatic red hexagons) looks less likely because the cyclization of the three nonaromatic blue hexagons must overcome very high barriers, as shown in fig. S23, which might explain why the hypothetical intermediate 6 in Fig. 2 was not observed in the mixture of ring closure products.

Spectrometric and electrochemical properties of 1b

1b was dissolved in CH2Cl2 to give an orange solution with two intense absorptions around 287 and 456 nm, as well as some weaker peaks. The longest wavelength absorption maximum appears at 526 nm, with a tail in the visible region that extends beyond 600 nm (Fig. 4A). Time-dependent (TD)–DFT calculations at the B3LYP/6-31G(d,p) level reveal that the transition energies and oscillator strengths show good agreement with the observed ultraviolet-visible (UV-vis) absorption values of 1b (Fig. 4A). The attributions of major absorption bands at 292, 350, and 455 nm are listed in table S3. For comparison, spectra of both grossly warped nanographene (7; C80H30, with one pentagon and five heptagon defects; fig. S15) (17) and planar nanographene (8; C78H30, defect free; fig. S15) (18) are shown in fig. S24. The experimental absorption data confirm the DFT predictions that both 1a and 1b have slightly narrower bandgaps than those of 7 and 8 (EHOMO = −5.18, −5.10, −5.12, and −4.96 eV; ELUMO = −2.30, −2.25, −2.10, and −2.05 eV for 1a, 1b, 7, and 8, respectively) (HOMO indicates highest occupied molecular orbital; LUMO, lowest unoccupied molecular orbital) (figs. S25 to 27 and table S2). The emission of 1b was measured at ambient temperature with excitation at 456 nm, and fluorescence maxima were observed at ~565 and 604 nm with several additional shoulders (the measured quantum efficiency is about 3.3%) (Fig. 4A). The infrared (IR) absorption spectrum of 1b measured in the solid state, shown in Fig. 4B, covers the range from 450 to 3350 cm−1. It is dominated by three bands and two humps, labeled with asterisks, at 1120.5, 1384.7, 1629.6, 2850.4, and 2919.8 cm−1. A calculated spectrum is displayed in the same figure for comparison, and calculated positions of the most prominent bands are given in table S4.

Fig. 4 Spectrometric and electrochemical properties of the penta-mesityl carboncone[1,2] derivative (1b).

(A) Ultraviolet-visible (UV-vis) absorption (black solid line), the oscillator strengths (blue bars) obtained by TD-DFT calculations at the B3LYP/6-31G(d,p) level of theory, and emission spectra (red dashed line) of 1b in CH2Cl2. Inset: Digital photograph of a CH2Cl2 solution of 1b irradiated at 456 nm. (B) IR absorption spectrum of 1b. Asterisks mark the dominant bands and humps at 1120.5, 1384.7, 1637.3, 2850.4, and 2919.8 cm−1 in the observed spectrum (red). A theoretical spectrum (black) calculated at the B3LYP/6-31G(d,p) level of theory is shown for comparison. (C) Cyclic voltammogram of 1b and corannulene in N,N-dimethylformamide. (D) The first redox process of 1b measured in o-dichlorobenzene at variable scan rates (120, 100, 80, 60, 40, 20, and 10 mV/s).

Besides the spectrometric characterizations, the electrochemical properties of 1b were studied by cyclic voltammetry (Fig. 4, C and D). For comparison, the cyclic voltammogram of corannulene 2 was measured under the same conditions. The electrochemical data reveal that 1b has a much higher reduction voltage relative to that of 2 (Fig. 4C).

DISCUSSION

Synthesis of carboncone[1,2] under mild conditions

It is remarkable that the carboncone[1,2] derivatives 1 can be synthesized in just three simple steps, starting from commercially available corannulene 2. In particular, in the final organic chemical reaction that produces carboncone, the oxidative cyclizations do not stop at the fivefold annulated product pentaphenalenocorannulene 6, but they continue all the way to a completely closed product 1 by stitching together the five appendages of 5 (Fig. 2). The 10 new Csp2─Csp2 bonds are formed effectively to overcome the structural strain in the final closure step, which is quite mild, occurring even at 0°C. By contrast, the syntheses of buckybowls (11, 1314, 1920) with curvatures similar to carboncone[1,2] required much more severe reaction conditions, such as 180°C in solution or temperatures around 1000°C in the gas phase.

Within the toolbox of methods for critical constructions of Csp2─Csp2 bonds in strained hydrocarbon compounds, flash vacuum pyrolysis (FVP) (11) and metal-catalyzed intramolecular ring-closing arylations (1920) have been the two main strategies. Other strategies include reductive coupling followed by oxidation with DDQ (2,3-dichloro-5,6-dicyano-1,4-benzoquinone) (21) and C─F activation with silyl cation reported by Siegel and co-workers (22), McMurry coupling reported by Sygula and Rabideau (23), and alumina-promoted cove-region closure by elimination of hydrogen fluoride (HF) reported by Amsharov et al. (24). Recently, the venerable Scholl reaction, which has been pushed to spectacular heights by Müllen and co-workers (18) for the synthesis of many strain-free planar nanographenes in the last 10 to 15 years, was used by Itami and co-workers (17) to synthesize grossly warped nanographenes (7; fig. S15). In contrast to those warped nanographenes (17, 25), which embody negative curvature owing to the presence of five heptagons in their structure, the present new carboncone[1,2] displays only positive curvature resulting from one pentagonal defect in the hexagonal lattice. To the best of our knowledge, the Scholl reaction has never been used before as a ring closure strategy to synthesize a buckybowl having positive curvature.

Electronic properties of the carboncone[1,2]

DFT calculations at the B3LYP/6-31G(d,p) level reveal that the LUMO of carboncone[1,2] (1) is lower lying than that of the parent corannulene 2 (figs. S25 and S26 and table S2), which is in agreement with the cyclic voltammetric data (Fig. 4C) with the much higher reduction voltage of 1b relative to that of 2. Likely carboncones with higher cone walls should have higher reduction voltages, corresponding to lower LUMO levels and stronger electron-accepting abilities. In addition, the scan rate–dependent potential difference of 1b (Fig. 4D) indicates a quasi-reversibility of the first redox process of 1b. It is likely that the remaining electrons are left on the pentagon of carboncone 1b, making it relatively harder to lose than the normal electron on the carboncone.

Bottom-up growth of carboncone[1,m] with carboncone[1,2] as seed

Benefiting from the bottom-up chemical synthesis strategies, some structurally well-defined carbon nanostructures, such as cycloparaphenylenes (26), the shortest CNT endcap (27), and carbon nanobelts (28), have been synthesized and sought as seeds for controlled syntheses of single-walled CNT (SWCNT) materials. A number of attempts to grow single-index SWCNTs from the well-defined seed molecules by analogous processes have already been reported (2934). The carbon skeleton of 1, with a single pentagon-containing cone cap encircled by a lattice of fused hexagons, corresponds to a 60° positive disclination defect introduced into a graphene network (Fig. 5) (1, 2). By the same token, this carboncone[1,2] molecule is very promising as a seed from which to grow structurally uniform carboncones by repetitive addition of carbon atoms to annulate new hexagonal rings (Fig. 5). All the expanded carboncones would have the same apex angles in the family of carboncone[1,m].

Fig. 5 Topology structure and strategy for the bottom-up synthesis of structurally uniform carboncones[1,m].

Carboncone[1,2] molecule (1) is formed by introducing a 60° positive disclination defect in the nanographene sheet, and the growth of carboncone[1,m > 2] undergoes epitaxial expansion from the carboncone[1,2] seed molecule (1).

CONCLUSIONS

In summary, the first carboncone[1,2] molecule 1a (C70H20) and its penta-mesityl derivative 1b have been effectively synthesized starting from corannulene C20H10 in just three steps using mild organic chemical reactions. The reaction mechanism for the feasible Scholl reaction in the critical ring closure process leading to the strained carboncone[1,2] molecule has been explained theoretically. The unique structure of the penta-mesityl carboncone 1b caused by one pentagon in the hexagonal lattice of trigonal carbon atoms was confirmed by single-crystal x-ray crystallography. DFT calculations reveal that the cone-shaped carboncone[1,2] is rigid and protected from cone-to-cone inversion by an exceptionally high energy barrier. The UV-vis and IR absorptions of 1b are shown to be comparable with those of sp2-hybridized planar and warped nanographene with similar size, whereas the electrochemical properties differentiate it from its parent corannulene. Carboncone[1,2] not only provides valuable clues into how other carboncone[n,m] might be prepared but also gives a very promising seed from which to controllably grow uniform carboncone[1,m] materials and a possibly suitable species for the study of carbon-rich molecules similar to fullerenes and PAHs in interstellar space.

MATERIALS AND METHODS

Materials

Unless otherwise noted, all materials including dry solvents were obtained from commercial suppliers and used without further purification. Unless otherwise noted, all reactions were performed with dry solvents under an atmosphere of argon in dried glassware with standard vacuum-line techniques. All workup and purification procedures were carried out with reagent-grade solvents in air. Analytical thin-layer chromatography (TLC) was performed using E. Merck silica gel 60 F254–precoated plates (0.25 mm). Corannulene 2 was synthesized according to the reported procedure (35).

NMR spectroscopy

1H NMR (400, 500, and 600 MHz) and 13C NMR (100, 150, and 210 MHz) spectra were recorded using a Bruker AV 400 (600 or 850) spectrometer at 298 or 323 K. Chemical shifts were reported on the delta scale in parts per million relative to CHCl3 (δ = 7.26 ppm), CH2Cl2 (δ = 5.32 ppm), and C2H2Cl4 (δ = 6.00 ppm) for 1H NMR, and CDCl3 (δ = 77.0 ppm) and CD2Cl2 (δ = 55.0 ppm) for 13C NMR. The 1H NMR spectra data were presented as follows: chemical shift, multiplicity (s, singlet; d, double; and m, multiplet), and integration.

Mass spectrometry

High-resolution mass spectra (HRMS) were determined on a Bruker Daltonics Ultraflex III Time-of-Flight (TOF)/TOF (MALDI-TOF MS) or on a JEOL JMS-S3000 Spiral TOF (MALDI-TOF MS).

IR spectroscopy

The IR spectra were obtained using a Thermo Scientific Nicolet 380 IR spectrometer.

UV-vis absorption and fluorescence emission spectroscopy

UV-vis absorption spectra were recorded on a Shimadzu UV-2550 spectrometer with a resolution of 0.5 nm. Emission spectra were recorded on an Edinburgh fls980 spectrometer with a resolution of 0.4 nm and automatically corrected by instrumental function. Dilute solutions in degassed spectral grade dichloromethane in a 1-cm square quartz cell were used for measurements. Absolute fluorescence quantum yields were determined with a Hamamatsu C9920-02 calibrated integrating sphere system equipped with a multichannel spectrometer (PMA-11).

Cyclic voltammetry

Cyclic voltammetry was performed on an Als/600D electrochemical analyzer. The cyclic voltammetric cell consisted of a Pt electrode, a Pt wire counter electrode, and an Ag/AgNO3 reference electrode. The measurements were carried out at a scan rate of 100 mV/s or a variable of 10 to 120 mV/s with tetrabutylammonium hexafluorophosphate (nBu4NPF6) as the supporting electrolyte.

X-ray diffraction

X-ray crystallographic data of compound 1b were collected on a Rigaku x-ray single crystal diffractometer (XtaLAB PRO MM007-DW) using a rotating-anode x-ray as the diffraction source with Cu Kα radiation (λ = 1.54184 Å) at 100 K. The structure was solved by intrinsic phasing with the program of SHELXT-2014 (36), and all nonhydrogen atoms were refined anisotropically by the full-matrix least-squares method on F2 with the program of SHELXL-2018 (36) within the program of Olex2 (37). The refined structure shows a slight disorder. The minor part of the central C70 with an occupancy of 0.328 was regulated using 46 restraints, whereas the major part with an occupancy of 0.672 was completely free of restraints with unambiguous geometric configuration and reasonable atomic temperature factors (the equivalent isotropic atomic displacement parameters Ueq ranging from 0.046 to 0.068 Å2). The obtained crystallographic data support that the structure had relatively separated electron density centers for each of the two disordered parts (with one-third and two-third occupancy, respectively), and the disorder had negligible influence to the determination of the structure in the major part with two-third occupancy.

Computational methods

DFT calculations were performed on the Gaussian 09 program (38) using the B3LYP method and 6-31G(d,p) basis set. The Cartesian coordinates and energies of all optimized species are shown in table S2. All transition states were obtained with one imaginary frequency and checked by intrinsic reaction coordinate (IRC) calculations. The mesityl groups were replaced by hydrogen atoms to reduce the workload of computation in the calculations of transition states in the dication pathway. The hexa-(4-mesityl-naphthyl)-coronene (9b) and its ring closure compound (10b) can be taken as the strain-free references (fig. S15) in the related strain energy calculations of 1b and 5b. The strain energies (∆G) of 1b and 5b were estimated by comparing the electronic and thermal free energies (G) of 1b and 5b to 5/6 of the corresponding strain-free reference compound 10b and 9b, respectively. In the calculation of cone-to-cone inversion energies, the S-shaped transition state structures of 1a and 1b are analogous to the transition state of the bowl-shaped fragment of C70 reported by Wu et al. (39) (shown in fig. S13 and table S2). In addition, the C─C bonds of the outer rings in the flat structure of 1a suffer so much stretching force that it will squeeze the flat structure into an S-shaped structure in the transition state calculation. To solve this dilemma, all carbon atoms were restricted in the same plane in the optimization of the flat structure for the transition state of 1a.

Synthesis of 4-mesitylnaphthalen-1-amine (S2)

To a solution of Pd(OAc)2 [112 mg, 0.5 mmol, and 5 mole percent (mol %)], SPhos (40) (400 mg, 1 mmol, and 0.1 equiv.), 4-bromonaphthalen-1-amine (S1, 2.21 g, 10 mmol, and 1.0 equiv.), and mesitylboronic acid (1.8 g, 11 mmol, and 1.1 equiv.) in toluene (100 ml) was added a solution of K3PO4 (6.36 g, 30 mmol, and 3.0 equiv.) in H2O (10 ml), and the resultant mixture was stirred at 100°C for 6 hours under nitrogen. After cooling the mixture to room temperature, the reaction mixture was extracted with CH2Cl2. The combined organic layers were dried over Na2SO4, and the solvent was removed under reduced pressure. The crude product was purified by silica gel column chromatography (eluent: hexane/CH2Cl2 = 1:2) to give 4-mesitylnaphthalen-1-amine (S2, 2.35 g, 90%) as a purple solid. 1H NMR (400 MHz, CDCl3) δ = 7.97 (d, J = 8.4 Hz, 1H), 7.54 (ddd, J = 8.3, 5.8, 2.3 Hz, 1H), 7.47 to 7.40 (m, 2H), 7.17 (d, J = 7.5 Hz, 1H), 7.11 (s, 2H), 6.94 (d, J = 7.5 Hz, 1H). 4.19 (s, 2H), 2.49 (s, 3H), and 2.01 (s, 6H). 13C NMR (100 MHz, CDCl3) δ = 141.13, 137.60, 137.21, 136.58, 132.71, 129.78, 128.08, 127.14, 126.26, 126.00, 124.79, 124.05, 121.19, 109.92, 21.22, 20.48, and 17.83. HRMS; mass/charge ratio (m/z) for C19H19N [M + H]+ calcd.: 262.1590, found 262.1594.

Synthesis of 1-bromo-4-mesitylnaphthalene (4b)

To a solution of 4-mesitylnaphthalen-1-amine (S2, 2.09 g, 8 mmol, 1 equiv.), nBu4NBr (5.15 g, 16 mmol, and 2 equiv.), and p-TsOH (1.65 g, 9.6 mmol, and 1.2 equiv.) in CH3CN (100 ml) in a 250-ml round bottom flask was injected 0.88 ml of nBuONO (1.0 g, 9.6 mmol, and 1.2 equiv.) with stirring. After 5 min, CuBr2 (90 mg, 0.40 mmol, and 0.05 equiv.) was added, and the resultant mixture was stirred at ambient temperature for 5 hours; the reaction progress was monitored by TLC. After completion of the reaction, the reaction mixture was extracted with CH2Cl2. The combined organic layers were dried over Na2SO4, and the solvent was removed under reduced pressure. The crude product was purified by silica gel column chromatography using pure hexane to give 1-bromo-4-mesitylnaphthalene (4b, 1.84 g, 71%) as a white solid. 1H NMR (400 MHz, CDCl3) δ = 8.38 (d, J = 8.5 Hz, 1H), 7.91 (d, J = 7.5 Hz, 1H), 7.64 (ddd, J = 8.2, 5.3, 2.5 Hz, 1H), 7.50 to 7.41 (m, 2H), 7.18 (d, J = 7.5 Hz, 1H), 7.08 (s, 2H). 2.46 (s, 3H), and 1.93 (s, 6H). 13CNMR (100 MHz, CDCl3) δ =139.08, 137.23, 136.78, 135.91, 133.33, 132.25, 129.90, 128.23, 127.55, 127.28, 127.20, 126.92, 126.08, 121.82, 21.20, and 20.36. HRMS; m/z for C19H17Br [M − H] calcd.: 323.0430 (100%) and 325.0410 (99%). found 323.0435 (100%) and 325.0406 (99%).

Synthesis of 1,3,5,7,9-pentakis(naphthyl)corannulene (5a)

To a solution of Pd(PPh3)4 (30 mg, 25 μmol, and 0.25 equiv.) in toluene (15 ml) were added 1,3,5,7,9-pentakis(Bpin)corannulene (3, 88 mg, 0.1 mmol, and 1 equiv.), 1-bromo-naphthalene (4a, 210 mg, 1.0 mmol, and 10 equiv.), Pd(PPh3)4 (30 mg, 25 μmol, and 0.25 equiv.), and a solution of Cs2CO3 (0.98 g, 3.0 mmol, and 30 equiv.) in water (5 ml), and the resultant mixture was stirred at 100°C for 24 hours under nitrogen. After cooling the mixture to room temperature, the reaction was quenched by dilute hydrochloric acid, and the mixture was extracted with CH2Cl2 and washed with saturated NaCl solution three times. The combined organic layers were dried over Na2SO4, and the solvent was removed under reduced pressure. The crude product was purified by silica gel column chromatography (eluent: hexane/CS2 = 1:2) to give 1,3,5,7,9-pentakis(naphthyl)corannulene (5a, 39 mg, 44.3%) as a white solid. The 1H NMR spectrum of 5a shows broad peaks due to slow rotations of the naphthyl groups at room temperature, and the spectroscopic data matched those reported in the literature (41). 1H NMR (400 MHz, CD2Cl2, 25°C) δ = 8.54 to 7.92 (m, 5H), 7.87 (m, 10H), and 7.59 to 7.26 (m, 25H). 1H NMR (400 MHz, C2D2Cl4, 80°C) δ = 7.92 to 7.86 (m, 10H), 7.57 (s, 5H), and 7.43 to 7.37 (m, 25H). HRMS (FT-ICR); m/z for C70H40 [M + H]+ calcd.: 880.3208, found 880.3220.

Synthesis of 1,3,5,7,9-pentakis(4-mesitylnaphthyl)corannulene (5b)

To a solution of Pd(PPh3)4 (44 mg, 37.5 μmol, and 0.25 equiv.) in toluene (20 ml) were added 1,3,5,7,9-pentakis(Bpin)corannulene (3, 132 mg, 0.15 mmol, and 1.0 equiv.), 1-bromo-4-mesitylnaphthalene (4b, 490 mg, 1.5 mmol, and 10.0 equiv.), and a solution of Cs2CO3 (1.47 g, 4.5 mmol, and 30.0 equiv.) in water (10 ml), and the resultant mixture was stirred at 100°C for 24 hours under nitrogen. After cooling the mixture to room temperature, the reaction was quenched by dilute hydrochloric acid, and the mixture was extracted with CH2Cl2. The combined organic layers were washed three times with saturated NaCl solution and dried over Na2SO4, and the solvent was removed under reduced pressure. The crude product was purified by silica gel column chromatography (eluent: hexane/CH2Cl2 = 1:4) to give 1,3,5,7,9-pentakis(4-mesitylnaphthyl)corannulene (5b, 78 mg, 35%) as a light yellow solid. The 1H NMR of 5b also shows broad peaks due to slow rotations of the 4-mesitylnaphthyl groups at room temperature. 1H NMR (500 MHz, CD2Cl2, 25°C) δ = 8.58 to 8.07 (m, 5H), 7.78 (m, 5H), 7.65 to 7.54 (m, 5H), 7.50 to 7.24 (m, 20H), 7.01 (s, 10H), 2.42 (s, 15H), 1.92 (s, 15H), and 1.34 (s, 15H). 1H NMR (400 MHz, C2D2Cl4, 80°C) δ = 7.80 to 7.24 (m, 35H), 7.01 (s, 10H), 2.42 (s, 15H), and 1.89 (s, 30H). HRMS (FT-ICR); m/z for C115H90 [M + H]+ calcd.: 1471.7121, found 1471.7167.

Synthesis of carboncone (1a)

To a solution of 1,3,5,7,9-pentakis(naphthyl)corannulene (5a, 39 mg, 0.044 mmol, and 1.0 equiv.) in dry CH2Cl2 (10 ml) was added DDQ (100 mg, 0.44 mmol, and 10.0 equiv.) at 25° or 0°C under nitrogen gas flow. After stirring for about 5 min, trifluoromethanesulfonic acid (0.1 ml) was added to the mixture. The mixture was further stirred for 10 min at 25° or 0°C. The reaction mixture was quenched by adding saturated aqueous NaHCO3. A black precipitate was formed and filtrated, which is almost insoluble in common organic solvents. Therefore, the crude product was only characterized by high-resolution MALDI-TOF MS, which indicates that the product is a mixture. The strongest signal corresponds to the desired carboncone 1a, and the other signals mainly belong to incomplete ring closure products. HRMS (MALDI-TOF); m/z for C70H20 [M]+• calcd.: 860.1559, found 860.1597.

Synthesis of carboncone (1b)

To a solution of 1,3,5,7,9-pentakis(4-mesitylnaphthyl)corannulene (5b, 30 mg, 0.02 mmol, and 1 equiv.) in dry CH2Cl2 (10 ml) was added DDQ (47 mg, 0.2 mmol, and 10 equiv.) at 25° or 0°C under nitrogen gas flow. After stirring for about 5 min, trifluoromethanesulfonic acid (0.1 ml) was added to the mixture. The mixture was further stirred for 10 min at 25° or 0°C. The reaction mixture was quenched by adding saturated aqueous NaHCO3, and then the mixture was extracted with CH2Cl2. The combined organic layer was dried over Na2SO4, and the solvent was removed under reduced pressure. The crude product was purified by silica gel column chromatography (eluent: hexane/CS2 = 1: 1) to give the carboncone (1b, 14 mg, 46%) as an orange solid. 1H NMR (500 MHz, CD2Cl2) δ = 9.91 (d, J = 9.6 Hz, 1H), 9.84 (s, 1H), 8.72 (d, J = 8.9 Hz, 1H), 7.43 (s, 1H), 7.14 (s, 1H), 2.70 (s, 3H), 2.61 (s, 3H), and 1.54 (s, 3H). 13C NMR (150 MHz, CD2Cl2 and CS2) δ = 192.60, 138.80, 138.71, 137.84, 137.47, 137.28, 136.97, 130.05, 129.02, 128.93, 128.65, 127.98, 127.91, 125.18, 125.08, 124.73, 124.68, 124.57, 123.72, 123.66, 123.53, 21.80, 21.71, and 20.61. HRMS (MALDI-TOF); m/z for C115H70 [M]+• calcd.: 1450.5472, found 1450.5450.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/8/eaaw0982/DC1

Fig. S1. Full synthetic scheme of 1.

Fig. S2. NMR spectra of S2.

Fig. S3. NMR spectra of 4b.

Fig. S4. 1H NMR (400 MHz) spectra of 5a.

Fig. S5. 1H NMR spectra of 5b.

Fig. S6. 13C NMR spectra of 5b.

Fig. S7. MALDI-TOF mass spectra of C70H20 (1a).

Fig. S8. MALDI-TOF mass spectra of C115H70 (1b).

Fig. S9. 1H NMR spectrum of the carboncone (1b).

Fig. S10. Theoretical simulations for 1H NMR spectra of 1b and the possible heptagon-containing isomer.

Fig. S11. 13C NMR spectrum of the carboncone (1b).

Fig. S12. Packing mode of 1b in the crystal.

Fig. S13. The transition state (TS) structures for cone-to-cone inversion of 1a and 1b.

Fig. S14. IRC calculation results of the inversion TS of 1a and 1b.

Fig. S15. Structures of compounds 7, 8, 9b, and 10b.

Fig. S16. Energy barriers for different ring closure steps on A2 and the relative energies of the intermediate products by a dication pathway.

Fig. S17. Energy barriers for different ring closure steps on A3 and the relative energies of the intermediate products by a dication pathway.

Fig. S18. Energy barriers for different ring closure steps on A4 and the relative energies of the intermediate products by a dication pathway.

Fig. S19. Energy barriers for different ring closure steps on A5 and the relative energies of the intermediate products by a dication pathway.

Fig. S20. Energy barriers for different ring closure steps on A6 and the relative energies of the intermediate products by a dication pathway.

Fig. S21. Energy barriers for different ring closure steps on A7 and the relative energies of the intermediate products by a dication pathway.

Fig. S22. Energy barriers for different ring closure steps on A8 and the relative energies of the intermediate products by a dication pathway.

Fig. S23. Energy barriers for different ring closure steps on A3′ and the relative energies of the intermediate products by a dication pathway.

Fig. S24. UV-vis absorption spectra of 7 and 8.

Fig. S25. Molecular orbitals (from HOMO−1 to LUMO+1) of carboncone 1a and 1b calculated at the B3LYP/6-31G(d,p) level.

Fig. S26. Molecular orbitals (from HOMO−1 to LUMO+1) of nanographene 7 calculated at the B3LYP/6-31G(d,p) level.

Fig. S27. Molecular orbitals (from HOMO−1 to LUMO+1) of 2 and 8 calculated at the B3LYP/6-31G(d,p) level.

Table S1. Crystallographic data and structure refinement details of 1b.

Table S2. Cartesian coordinates of optimized species at the B3LYP/6-31G(d,p) level on Gaussian 09.

Table S3. UV-vis absorption of 1a predicted by TD-DFT calculations at the B3LYP/6-31G(d,p) level.

Table S4. Calculated vibrational frequencies at the B3LYP/6-31G(d,p) level for prominent bands of carboncone 1b in the IR spectrum (in cm−1).

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: We thank K. Shi for the constructive discussion on CV measurement, Z.-W. Lin and H.-F. Su for the MS measurement, D. Lin and L. Feng from the High-Field NMR Center of Xiamen University for the 850-MHz NMR measurement, and S.-Z. Zhan from Shantou University for assistance during crystallographic data collection. Funding: This research was supported by the National Natural Science Foundation of China (21771152, 21721001, 21390390, 21827801, 51572231, 21571151, and 21701134), the 973 Program of China (2015CB932301), the Major Science and Technology Project between University-Industry Cooperation in Fujian Province (2016H6023), and the Fundamental Research Funds for the Central Universities (20720170028 and 20720160084). This research was also supported financially by the U.S. National Science Foundation (CHE-0809494 and CHE-1149096). Author contributions: Z.-Z.Z. synthesized the materials, conducted the NMR and optical measurements, prepared the crystals, and wrote the manuscript draft. Z.-C.C. performed the theoretical calculations and analyzed the theoretical data. C.-H.C. conducted the cyclic voltammetry experiment. Q.Z. designed the experiment and wrote the manuscript. S.-Y.X. organized the research and revised the manuscript. J.M.Q. participated in the synthesis at the early stage. L.T.S. conceived the early concept of carboncone and revised the manuscript. S.-H.L. provided help in the cyclic voltammetry experiment and NMR spectroscopy, and plotted some of the figures for the manuscript and the Supplementary Materials. H.-R.T., Y.-R.Y., S.-H.L., and F.-F.X. conducted the SCXRD characterization and solved the crystallographic structure. X.-J.Z. and P.-Y.X. explored the early route for synthesis of mesityl-substituted derivative. X.-M.X. supplemented the experimental data in the revision. S.-L.D., Y.-Z.T., R.-B.H., and L.-S.Z. participated in the discussion and analysis of the data. All the authors discussed the experiment and commented on the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors. The crystallographic data for the structure reported in this paper have been deposited at the Cambridge Crystallographic Data Centre (CCDC) with deposition number CCDC 1846302 [1b, C115H70]. Copies of the data can be obtained free of charge from www.ccdc.cam.ac.uk/data_request/cif.
View Abstract

Navigate This Article