A saddle-shaped OBO-doped nanographene: facile synthesis, adaptive double-layer assembly, and enhanced Lewis acidity

Abstract

Current OBO-doped nanographenes (NGs) are predominantly planar, and the precise synthesis of curved OBO-doped NGs remains a challenge owing to their limited thermodynamic stability. Herein, we report a two-step synthesis of the first saddle-shaped OBO-doped NGs (1a–1c) built upon a corannulene core. Pristine 1a exhibits reversible thermo-responsive behavior characteristic of a lower critical solution temperature (LCST)-type response. Single-crystal X-ray diffraction analysis combined with theoretical calculations reveals that these nanographenes adopt a saddle geometry. Notably, the incorporation of bulky triisopropylphenyl (TIPP) groups not only improves thermodynamic stability but also enables 1c to form ordered, adaptive bilayer assemblies in the solid state, inducing dramatic flattening of the corannulene core—a rare example of structure adaptation driven by supramolecular assembly rather than intramolecular steric effects. Crystallographic analyses further demonstrate that such geometric distortion increases the accessibility of boron centers. Consequently, the saddle-shaped topology imparts significantly enhanced Lewis acidity to these NGs, facilitating the first crystallographically characterized dual methanol coordination on the concave surface of a boron-doped nanographene. This work introduces a unique topological scaffold combining non-planarity, multi-heteroatom doping, and bilayer assembly, expanding the family of heteroatom-doped NGs with novel topologies.

---

Introduction

Nanographenes (NGs), as two-dimensional π-conjugated materials with tunable electronic and optical properties, have garnered immense attention in the past decades for their potential applications in field-effect transistors, light-emitting diodes, solar cells and electrocatalysts.^1–5 The structural topology of NGs—ranging from planar to curved (bowl-shaped, saddle-shaped, warped, or tubular)—plays a pivotal role in determining their physical and chemical behaviors, as curvature introduces intrinsic electronic polarization and modulates intermolecular interactions.^1,6–8 Corannulene, a classic curved polycyclic aromatic hydrocarbon (PAH) with a bowl-shaped structure, serves as an ideal building block for constructing curved NGs due to its inherent C₅ symmetry and ability to undergo regioselective functionalization.^9–17 In this field, chemists like Scott,¹² Siegel,¹³ and Itami^14,15 have made great contributions. However, the synthesis of curved NGs with well-defined non-hexagonal rings still remains a significant challenge, as the introduction of strained rings often leads to structural instability and synthetic uncontrollability.^18–23

Heteroatom doping has proven particularly effective for functionalizing NGs, offering enhanced stability and tunable electronic properties for organic electronic applications.^24–27 This approach has recently enabled the solution-phase synthesis of the largest NBN-doped triangulene by Miao and colleagues in 2021.²⁸ Recently, Liu and co-workers reported NBN-edged porous nanographenes with unusual frontier molecular orbital distribution.²⁹ Meanwhile, OBO doping, in which B and O atoms form three-membered O–B–O linkages or integrated heterocyclic rings, has emerged as a promising approach for enhancing the stability and tuning the aromaticity, Lewis acidity, and intermolecular assembly of NGs.^30–34 For instance, in 2016, Müllen, Feng, Wang and co-workers reported an OBO-doped bistetracene analogue with stable zigzag edges.³⁰ Wang and Yang independently developed OBO-doped triangulene analogues in 2022 (ref. 31 and 32) (Fig. 1). Notably, these previous studies on OBO-doped NGs have primarily focused on planar structures with low-strain six-membered rings, while saddle-shaped OBO-doped NGs with unique symmetry have rarely been reported. The incorporation of non-hexagonal rings (e.g., pentagons and heptagons) into OBO-doped NGs is of particular interest, as these rings induce curvature and create distinct electronic environments, yet their synthesis is hampered by the high strain energy associated with ring fusion.^35,36

Fig. 1 Representative planar NBN-doped/OBO-doped NGs and novel saddle-shaped OBO-doped NGs in this work.

Another critical aspect of curved NGs is their self-assembly behavior. Due to the large contact area on both sides of the π-core, nanographenes typically tend to aggregate, driven by dispersion forces.^37–40 Recently, self-assembly of nanographenes with a defined number of layers has garnered widespread interest.^41–43 However, the multilayer self-assembly of saddle-shaped NGs remains rather unexplored.^44,45 Controlling the self-assembly of saddle-shaped NGs into well-defined superstructures is essential for optimizing their collective properties, such as charge transport and light harvesting.⁴⁶ Additionally, the Lewis acidity of B-containing NGs enables coordination with nucleophiles (e.g., F⁻ and amines),^47–49 but coordination with weakly basic molecules like methanol has not been documented, representing an unexploited avenue to modulate their structural and optical properties.

Building on the versatile sym-1,3,5,7,9-pentakis(Bpin)corannulene platform,⁵⁰ we herein report the synthesis of a unique family of saddle-shaped OBO-doped NGs (1a–1c) via a concise two-step route. These nanographenes feature a corannulene core fused with five heptagonal rings and five OBO-doped heterocyclic segments, representing the first example of B/O-doped NGs integrating five seven-membered rings. Pristine 1a exhibits reversible thermo-responsive behavior. By introducing sterically demanding solubilizing groups (mesityl for 1b and TIPP for 1c), we achieved improved solubility and stability, thereby enabling the unambiguous structural characterization of a saddle-shaped OBO-doped NG. Remarkably, the NG (1c) undergoes adaptive self-assembly into a staggered bilayer structure, and its enhanced Lewis acidity allows for unique methanol coordination. This work not only enriches the synthetic methodology for curved heteroatom-doped NGs but also uncovers novel structure–property relationships, paving the way for the rational design of functional π-conjugated materials.

Results and discussion

Synthesis, purification and structural characterization of saddle-shaped OBO-doped nanographenes

Starting from the sym-1,3,5,7,9-pentakis(Bpin)corannulene platform reported by Scott et al. (2012),⁵⁰ we developed a concise two-step route to OBO-doped nanographenes 1a–1c. As depicted in Fig. 2a, the OBO-doped nanographenes were synthesized via two principal steps. First, a five-fold coupling reaction between sym-1,3,5,7,9-pentakis(Bpin)corannulene and tailored aryl partners afforded cyclization precursors 2a, 2b, and 2c in isolated yields of 72%, 66%, and 48%, respectively. Subsequently, BBr₃-mediated electrophilic borylation yielded target compounds 1a, 1b, and 1c in yields of 52%, 23%, and 78%, respectively. The structures and symmetry of 1a, 1b, and 1c were established by ¹H NMR spectroscopy (Fig. S25–S39). All three compounds exhibit time-averaged C₅ symmetry in solution on the NMR timescale. For instance, 1c displays two well-resolved singlets of equal intensity in the low-field region (Fig. 2b), consistent with a highly symmetric architecture. The signal at 7.06 ppm is assigned to triisopropylphenyl protons (Hc), while the resonance at 7.18 ppm corresponds to skeletal protons Ha and Hb; the low-field shift of the latter arises from the extended conjugation effect within the scaffold. High-resolution matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) confirmed the molecular compositions of 1a, 1b, and 1c (Fig. 2c and S20–S22), with observed m/z values matching simulated spectra. 1a (calcd for C₅₀H₁₅B₅O₁₀ [M]: 830.1125, found: 830.1154); 1b (calcd for C₉₅H₆₅B₅O₁₀ [M]: 1420.4988, found: 1420.5022); and 1c (calcd for C₁₂₅H₁₂₅B₅O₁₀ [M]: 1841.0157, found: 1841.0182). Notably, an additional signal at m/z = 3862.0360 was observed for 1c, consistent with the formation of a dimeric aggregate in the gas phase (Fig. 2c).

Fig. 2 Synthesis route, mass spectrometry and ¹H NMR spectrum of 1c. (a) Reaction conditions: (i) for 2: X-Phos, Pd(PPh₃)₄, Cs₂CO₃, DMF, 85 °C (2a)/130 °C (2b and 2c), 12 h; (ii) for 1a–1c: BBr₃, 150 °C, 24 h. (b) The ¹H NMR spectrum of 1c. (c) The MALDI-TOF MS of 1c.

Compound 1a exhibits poor solubility in low-to-moderate polarity solvents (e.g. dichloromethane and toluene) but good solubility in highly polar solvents such as DMF and DMSO. This behavior can be rationalized by strong Lewis acid–base interactions between the boron centers and the polar donor solvents. Consistent with this picture, ¹¹B NMR spectra of 1b and 1c in DMSO-d₆ exhibit pronounced high-field shifts relative to those in CDCl₃ (Fig. S34, S35, S40 and S41), indicating the formation of tetracoordinate boron–solvent adducts. Similar solvent-dependent trends were observed in the absorption spectra of 1b and 1c (Fig. S44 and S45). Interestingly, heating the DMF solution of 1a induced gradual precipitation, which reversed upon cooling—a thermo-reversible phase transition confirmed by variable-temperature ¹H NMR and dynamic light scattering (Fig. S28 and S29). This behavior is characteristic of lower critical solution temperature (LCST)-type response,⁵¹ attributed to thermally weakened Lewis acid–base interactions between 1a and DMF, allowing intermolecular π–π stacking to dominate and drive aggregation.

To improve the solubility in low-polar solvents, mesityl groups were incorporated to afford 1b, which indeed showed enhanced solubility in solvents such as dichloromethane and toluene. However, silica gel chromatography purification of 1b led to a significant yield loss (23%), likely due to its sensitivity toward weakly acidic surface sites. To address this stability issue, more sterically hindered triisopropylphenyl (TIPP) groups were installed to afford 1c, which could be purified smoothly by silica gel chromatography in a high yield of 78%. The improved stability and processability of 1c arise from the increased hydrophobic effect provided by the bulky isopropyl groups. Long-term solution stability studies further confirmed that 1c is significantly more robust than 1b.

Given the potential structural strain imposed by the combination of non-hexagonal rings and OBO doping, single crystals of 1c suitable for single-crystal X-ray diffraction (SCXRD) analysis were first grown via vapor diffusion (toluene/CHCl₃, 60 °C). The solid-state structure confirms a weak saddle-shaped topology composed of a central corannulene core surrounded by five heptagonal rings and five zigzag-edged segments. To the best of our knowledge, this represents the first example of a B/O-doped nanographene integrating five seven-membered rings. In contrast to planar OBO-doped nanographenes, 1c exhibits a shallow saddle geometry with a saddle depth of 1.22 Å (Fig. 3a), which is significantly shallower than those observed for the gas-phase monomer (2.73 Å, Fig. S61) and dimer (3.21 Å, Fig. 3b) of 1a. This phenomenon likely arises from intermolecular interaction forces among peripheral functional groups, which flatten the intrinsic saddle-shaped conformation characteristic of the unsubstituted skeletal framework. Remarkably, the crystal packing reveals a staggered bilayer arrangement (twist angle = 35.87°; Fig. S8), effectively minimizing repulsion between TIPP groups. Such well-defined bilayer stacking is relatively uncommon among saddle-shaped NGs compared to their planar analogues. Due to slight disorder in the core region, 1c was recrystallized via slow evaporation at 0 °C, yielding highly ordered crystals (Fig. 3a). The refined structure displays a discrete dimeric bilayer motif with a torsion angle of 34.37° and an interlayer distance of 3.44 Å (Fig. 3b), indicative of moderate π–π interactions. Similar to the previously reported multilayer stacking mode of nanographene,^52,53 this system rotates around the quasi-C₅ axis to accommodate the bulky groups. In addition, an inter-dimer distance is 6.95 Å, sufficiently large to suppress significant π–π interactions between adjacent dimers due to steric shielding.

Fig. 3 Single-crystal X-ray molecular structure and crystal packing arrangements of 1c grown at 0 °C, with the NICS(1) values of 1c. All hydrogen atoms and solvent molecules were omitted for clarity. (a) The single-molecule structure of 1c, the top view (thermal ellipsoids of 30% probability), side view, corannulene core and bond lengths of the corannulene rim in 1c. Oxygen and boron were marked in pink and blue, respectively. (b) The dimer structure of 1c, the top view (thermal ellipsoids of 30% probability) and the side view. Oxygen and boron were marked in pink and blue, respectively. Inset: the gas-phase structure of the dimer of 1a. (c) Molecular columnar packing arrangement (the triisopropylphenyl groups were omitted for clarity). (d) Bond lengths of the polycyclic framework of 1c. (e) NICS(1) values of 1c calculated at the B3LYP/6-31G(d,p) level.

Detailed structural analysis of this double-layer assembly revealed an unexpected flattening of the corannulene core in 1c. The distance between the pentagon center and the mean plane of the rim atoms is only 0.02 Å—dramatically smaller than that of pristine corannulene (0.87 Å)^54,55 and previously reported warped nanographenes (0.37 Å).¹⁵ In contrast to earlier examples of core flattening driven by intramolecular steric repulsion between peripheral substituent,^56,57 the planarization observed here is attributed to adaptive double-layer self-assembly. Density functional theory (DFT) calculations support this cooperative structural adaptation, revealing a simultaneous reduction in both bowl depth and saddle height upon dimerization (Fig. S64 and S65). Notably, such flattening is absent in the optimized structure of 1a (Fig. S62 and S63), highlighting the critical role of the bulky TIPP groups in directing the double-layer assembly. Furthermore, dimers of 1c exhibit columnar stacking along the c-axis, stabilized by inter-dimer C–H⋯π interactions (Fig. 3c).

Aromaticity was evaluated through bond-length analysis and NICS(1) calculations (Fig. 3d and e). The outer benzene rings (A₁–A₅) show bond-length equalization (1.369(7)–1.423(7) Å) and strongly negative NICS(1) values (−8.89 to −9.36), confirming their aromatic character. In contrast, all the heptagons and the hexagons constructed from OBO units are nonaromatic, as evidenced by the NICS(1) values: the BOC₅ rings exhibit slightly positive values, while the BOC₄ rings display only marginally negative values (Fig. 3e). The B–O bond lengths in BOC₄ (1.347(7)–1.361(7) Å) and BOC₅ rings (1.332(7)–1.347(7) Å) are shorter than those in OBO-doped planar nanographenes (1.378–1.440 Å)^31,32 and other OBO-doped PAHs.^30,58,59 This shortening results in enhanced π-donation from oxygen to boron. Elongated bonds in the corannulene rim (∼1.56 Å) relative to gas-phase optimized structures (∼1.44 Å) may result from planarization-induced strain. The anisotropy of the induced current density (ACID) calculations further supports these findings, revealing a pronounced clockwise diatropic ring current in rings A₁–A₅, indicative of their aromatic nature (Fig. S71). However, the current appears disrupted at the B–O bonds in the ACID plots, which does not meet the criteria for antiaromaticity. Therefore, the aromaticity of the BOC₅ ring is described as non-aromatic, which is also consistent with the NICS(1) calculation. In addition, as the OBO unit sites are located at the junction of the heptagon (BOC₅) and hexagon (BOC₄), the angle of ∠OBO is compressed to 110.01–111.61° (Table S2), smaller than the 120° required for ideal sp² hybridization. This geometric distortion disrupts the intrinsic three-center four-electron aromaticity of the OBO motif and renders the vacant boron orbitals more accessible for external Lewis base coordination.

Dimer interaction energy analysis

To quantify the non-covalent interactions governing dimer formation, fragmentation-based energy calculations⁶⁰ were performed on the theoretically optimized dimers of 1a, 1b and 1c. As shown in Fig. 4, the dimer binding energy of 1c (−127.13 kcal mol⁻¹) is significantly stronger than those of 1a (−62.13 kcal mol⁻¹) and 1b (−97.28 kcal mol⁻¹). Control analysis indicates that the π–π interactions between the conjugated backbones are comparable across the series, implying that the energetic difference arises mainly from substituent effects. Large dispersion interaction in XEDA II (Fig. S66) indicates that the dimer stability of 1c is dominated by London dispersion forces, originating from strong instantaneous dipole–dipole attractions between polarizable electron clouds, typical of nonpolar/weakly polar systems. The exceptional stability of the 1c dimer can be attributed to a synergistic combination of non-covalent forces: (i) π–π stacking between the curved conjugated backbones, (ii) C–H⋯π interactions among the bulky TIPP substituents, and (iii) C–H⋯O contacts between the alkyl substituents and the oxygen atoms of the OBO units.

Fig. 4 Interaction energy analysis for the dimers of 1 using fragmentation methods: (a) 1a dimer, (b) 1b dimer and (c) 1c dimer.

Optical properties and electronic structures of saddle-shaped OBO-doped nanographenes

The UV/vis absorption and fluorescence spectra of compounds 1b and 1c were recorded in dilute dichloromethane solutions. Both compounds exhibit well-resolved vibronic bands with prominent absorption features spanning 300–520 nm (Fig. 5a), demonstrating a pronounced red shift compared to sym-pentaphenylcorannulene⁵⁰ and related OBO-doped triangulenes,^31,32 consistent with an extended conjugation. Specifically, 1b shows a maximum absorption at 396 nm with shoulders at 442 and 463 nm, whereas 1c displays a slightly red-shifted profile with maxima at 399, 446, and 468 nm. This subtle difference can be attributed to the slightly stronger electron-donating effect of the TIPP groups in 1c relative to the mesityl groups in 1b, which subtly modulates the frontier orbital energies. Time-dependent density functional theory calculations reproduce the experimental trend, with transition energies of 3.28 eV for 1a, 3.24 eV for 1b and 3.23 eV for 1c (Fig. 5b), all lower than that of corannulene. The low-energy absorption bands are assigned to HOMO → LUMO transitions (Fig. S68–S70 and Tables S6–S8), consistent with the observed blue shift of 1b relative to 1c. Moreover, the exceptionally small oscillator strengths of 1a, 1b and 1c (0.0137, 0.0229 and 0.0254, respectively) further corroborate the weakly allowed and partially forbidden nature of their lowest excited states. The corresponding molar extinction coefficients (ε_max) were determined to be 7.21 × 10⁴ M⁻¹ cm⁻¹ for 1b and 7.35 × 10⁴ M⁻¹ cm⁻¹ for 1c. Emission maxima occurred at 563 nm (1b) and 513 nm (1c), with Stokes shifts of 100 nm and 45 nm, respectively. Transient photoluminescence analysis yielded quantum yields (Φ) and emission lifetimes (τ) as follows: 1a (Φ = 2.4%, τ = 3.1 ns), 1b (Φ = 11.5%, τ = 3.6 ns), and 1c (Φ = 14.4%, τ = 4.3 ns) (Fig. S49–S51).The fluorescence quantum yield of 1 is relatively low compared to those of OBO-doped triangulenes (31–58%),^31,32 likely due to enhanced vibrational relaxation in the flexible saddle-shaped framework.

Fig. 5 Optical spectra (recorded in dichloromethane, 10 µM) and molecular orbitals. (a) UV/vis spectra of 1b and 1c, with an inset photograph in the right corner displaying the emission spectra of 1b and 1c (represented by blue and green dotted lines, respectively). (b) Molecular orbitals of 1a, 1b and 1c.

The higher fluorescence quantum yield of 1c compared to 1b can be rationalized by the steric restriction imposed by the bulky TIPP groups, which hinder the fast rotation of peripheral aromatic rings. This restriction suppresses non-radiative decay (2.39 × 10⁸ s⁻¹ for 1b and 1.99 × 10⁸ s⁻¹ for 1c) and enhances radiative decay (3.10 × 10⁷ s⁻¹ for 1b and 3.35 × 10⁷ s⁻¹ for 1c, Table S4). In contrast, the mesityl groups in 1b rotate more freely. The constrained geometry also reduces structural relaxation upon excitation, minimizing the difference between ground- and excited-state geometries and thus reducing the Stokes shift. Electrochemically, both 1b and 1c exhibit quasi-reversible reduction waves in cyclic voltammetry (CV), with half-wave potentials (E^red1_1/2) at −1.42 V and −1.40 V (Fig. S52), respectively, in agreement with the trend in calculated LUMO energies.

Lewis acidity and dual methanol coordination

To evaluate the Lewis acidity and binding behaviors of boron atoms, titrations of 1b and 1c with tetrabutylammonium fluoride (TBAF) in tetrahydrofuran (THF) were performed, leading to dramatic changes in absorption and emission spectra. New absorption bands emerged at 352 and 382 nm for 1b and 341 and 375 nm for 1c, accompanied by overall blue shifts of the spectra (Fig. S53 and S54), along with distinct emission color changes (Fig. S58). However, non-linear curve-fitting analyses did not yield consistent or chemically reasonable binding stoichiometries (Table S5), suggesting that fluoride binding under these conditions does not follow a simple 1 : 1 or 1 : 2 model but involves more complex cooperative or multisite binding.

Intriguingly, slow diffusion of methanol vapor into a chloroform solution of 1c afforded single crystals of a well-defined methanol-coordinated adduct, 2CH₃OH·1c, which was fully characterized by SCXRD (Fig. 6a). The solid-state structure confirms tetracoordinate boron centers, with two methanol molecules selectively bound to two boron sites on the concave face of the saddle (Fig. S9). To our knowledge, this represents the first documented case of methanol coordination in a boron-doped nanographene. Although molecular dimensions remain similar to 1c, methanol binding induces a dramatic structural adaptation: the saddle height increases from 1.22 Å to 2.45 Å, and the corannulene core reverts to a distinct curved bowl geometry (depth = 0.59 Å). The B–O bond distances of 1.563(8) Å and 1.57(1) Å are characteristic of dative bonds, markedly longer than covalent B–O bonds in methoxide complexes and consistent with calculated values (Fig. S78). This coordination also reduces OBO bond angles (Table S2) and amplifies negative curvature of the heptagonal rings, with curvature parameters (k(v))¹⁸ approximately double (Fig. 6b and S19) those of fluoride-coordinated analogues,^41–43 indicating greater distortion. In the solid state, 2CH₃OH·1c forms a herringbone arrangement along the b-axis (layer spacing = 7.34 Å), stabilized by C–H⋯π interactions (Fig. 6c and S12).

Fig. 6 Single-crystal X-ray molecular structure, crystal packing arrangements and time-dependent UV/vis spectra of 2CH₃OH·1c. (a) The top view (thermal ellipsoids of 30% probability), side view of 2CH₃OH·1c, and the corannulene core in the crystal structure (thermal ellipsoids of 30% probability). (b) The methoxy-coordinated heptagons in the crystal structures, with the centroid shown as a green ball. (c) Crystal packing arrangements. All hydrogen atoms and solvent molecules were omitted for clarity. (d) Time-dependent UV/vis spectra of 1c (8.2 µM) in methanol solution. The zero-time point for spectral acquisition was defined as the moment to start testing immediately after dissolution.

Computational scanning of possible 2CH₃OH⋯1c and 4CH₃OH⋯1c isomers showed very small energy differences, suggesting the coexistence of multiple coordination isomers in solution. Notably, the crystallographically characterized isomer (2CH₃OH⋯1c-1) was identified as the most stable among the 2CH₃OH⋯1c series (Fig. S78). Furthermore, calculations simulating the stepwise addition of 1–5 methanol molecules to 1c indicate that stable coordination complexes are formed with three to four methanol molecules, as the incremental binding energy drops significantly for the fifth methanol (|ΔE_int(inc)| ≈ 5.04 kcal mol⁻¹; Fig. S77). Moreover, as the number of methanol molecules increases, the coordination bond lengths between methanol and 1c (r_O⋯B) gradually converge to the values observed in the crystallographic data (Fig. S80).

The dynamics of methanol coordination were probed experimentally. Titrations of 1c with methanol in dichloromethane, monitored by ¹H-NMR and UV/vis spectroscopy (Fig. S59 and S60), revealed negligible changes in chemical shifts and only minor spectral variations. However, time-dependent UV/vis spectra in pure methanol revealed a gradual blue shift over time (Fig. 6d), demonstrating that the coordination process is thermodynamically favorable but kinetically slow. DFT calculations of 1c with 1–5 methanol molecules in a solvation model reproduced this spectral blue-shifting trend (Fig. 6d and S76), validating the sequential coordination mechanism.

Finally, the Lewis acidity of 1c was compared with that of planar OBO-doped triangulene analogues^31,32 via coordination energy calculations (Fig. S81). The results clearly indicate enhanced Lewis acidity in saddle-shaped OBO-nanographene 1c, supported by its lowered LUMO level and stronger calculated binding energy toward methanol. This enhanced Lewis acidity can be attributed to the reduced steric shielding around the boron centers in the curved scaffold (Fig. S13, OBCC torsion angles in BOC₄ or BOC₅ rings: 0.05–17.51°), in contrast to the more congested environment in planar analogues (OBCC torsion angles in BOC₄ rings: 0.42–1.52°).^31,32

Conclusions

In summary, we have designed and synthesized a unique family of saddle-shaped OBO-doped nanographenes based on a corannulene core fused with five heptagonal rings, via a concise two-step synthetic strategy. Although DFT calculation calculations predict a typical saddle-shaped geometry, single-crystal X-ray diffraction analysis reveals that these curved nanographenes adopt a shallow saddle geometry and unexpectedly form ordered adaptive double-layer assemblies in the solid state. Notably, such bilayer packing induces a dramatic flattening of the corannulene core, which is distinct from previously reported sterically driven planarization and represents a rare example of structure adaptation driven by supramolecular assembly.

These saddle-shaped nanographenes exhibit structure-dependent solubility, thermo-responsive behavior, and tunable optical and electronic properties. The bulky triisopropylphenyl substituents significantly enhance stability, solubility, and dimer-binding energy through synergistic non-covalent interactions. Aromaticity analyses reveal that the outer benzene rings maintain strong aromaticity, while the OBO-containing heterocyclic units are rendered non-aromatic due to geometric distortion, which in turn increases the accessibility of the boron centers. Benefiting from the unique curved topology, these OBO-doped nanographenes show enhanced Lewis acidity and enable the rare crystallographic observation of dual methanol coordination on the concave surface. This work not only expands the structural diversity of heteroatom-doped curved nanocarbons but also establishes a direct link between three-dimensional molecular topology, adaptive supramolecular assembly, and Lewis acid functionality, offering a new design principle for functional polycyclic aromatic materials.

Author contributions

Q. Zhang and S.-Y. Xie conceived of the idea and directed the research. K. Zhang designed the protocols, performed the experiments, and analyzed the data. Z.-C. Chen and J. Li conducted density functional theory calculations. Q.-S. Deng and X.-M. Wu participated in organic synthesis. S. Jia, Y.-R. Yao and H.-R. Tian performed X-ray crystallography. K. Zhang, Q. Zhang and S. Jia wrote the manuscript. Q. Zhang, S.-Y. Xie, and S.-L. Deng coordinated and supervised the project. All authors discussed the results and revised the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

CCDC 2486275 (1c at 0 °C), 2486277 (1c at 60 °C) and 2486279 (2CH₃OH·1c at 0 °C) contain the supplementary crystallographic data for this paper.^61a–c

The experimental data and the characterization data are available in the supplementary information (SI). All relevant data that support the findings of this work are available from the corresponding author on request. The atomic coordinates of the optimized computational model data generated in this study are provided in the Source Data file. Source data are provided with this paper. Supplementary information is available. See DOI: https://doi.org/10.1039/d6sc02521h.

Acknowledgements

We thank the National Natural Science Foundation of China (NSFC) (22271238, 22571264, 92361303, 22305200, 22301249 and 21721001) for support.

References

1. A. Narita, X.-Y. Wang, X. Feng and K. Müllen, Chem. Soc. Rev., 2015, 44, 6616–6643.
2. B. Wu, H. Meng, X. Chen, Y. Guo, L. Jiang, X. Shi, J. Zhu, J. Long, W. Gao, F. Zeng, W.-J. Jiang, Y. Zhu, D. Wang and L. Mai, Angew. Chem., Int. Ed., 2025, 64, e202415071.
3. Z. Liu, S. Fu, X. Liu, A. Narita, P. Samorì, M. Bonn and H. I. Wang, Adv. Sci., 2022, 9, 2106055.
4. Y. Gu, Z. Qiu and K. Müllen, J. Am. Chem. Soc., 2022, 144, 11499–11524.
5. Y. Zhu and M. A. Petrukhina, Chem. Soc. Rev., 2026, 55, 3061–3077.
6. T. Enoki, Y. Kobayashi and K.-I. Fukui, Int. Rev. Phys. Chem., 2007, 26, 609–645.
7. P. Izquierdo-García, J. Lión-Villar, J. M. Fernández-García and N. Martín, Chem. Soc. Rev., 2025, 54, 11089–11104.
8. S. Mishra, D. Beyer, K. Eimre, S. Kezilebieke, R. Berger, O. Gröning, C. A. Pignedoli, K. Müllen, P. Liljeroth, P. Ruffieux, X. Feng and R. Fasel, Nat. Nanotechnol., 2020, 15, 22–28.
9. W. E. Barth and R. G. Lawton, J. Am. Chem. Soc., 1966, 88, 380.
10. X. Li, F. Kang and M. Inagaki, Small, 2016, 12, 3206.
11. M. C. Stuparu, Acc. Chem. Res., 2021, 54, 2858.
12. V. M. Tsefrikas and L. T. Scott, Chem. Rev., 2006, 106, 4868.
13. A. M. Butterfield, B. Gilomen and J. S. Siegel, Org. Process Res. Dev., 2012, 16, 664.
14. K. Kato, Y. Segawa, L. T. Scott and K. Itami, Angew. Chem., Int. Ed., 2018, 57, 1337–1341.
15. K. Kawasumi, Q. Zhang, Y. Segawa, L. T. Scott and K. Itami, Nat. Chem., 2013, 5, 739–744.
16. K. Shoyama and F. Würthner, J. Am. Chem. Soc., 2019, 141, 13008–13012.
17. Z.-Z. Zhu, Z.-C. Chen, Y.-R. Yao, C.-H. Cui, S.-H. Li, X.-J. Zhao, Q. Zhang, H.-R. Tian, P.-Y. Xu, F.-F. Xie, X.-M. Xie, Y.-Z. Tan, S.-L. Deng, J. M. Quimby, L. T. Scott, S.-Y. Xie, R.-B. Huang and L.-S. Zheng, Sci. Adv., 2019, 5, eaaw0982.
18. S. H. Pun and Q. Miao, Acc. Chem. Res., 2018, 51, 1630–1642.
19. W. Niu, J. Ma and X. Feng, Acc. Chem. Res., 2022, 55, 3322–3333.
20. Q. Zhang, T. Liu, Z. Gan, Y.-Z. Tan and S.-Y. Xie, Acc. Mater. Res., 2025, 6, 1476–1489.
21. Chaolumen, I. A. Stepek, K. E. Yamada, H. Ito and K. Itami, Angew. Chem., Int. Ed., 2021, 60, 23508–23532.
22. I. R. Márquez, S. Castro-Fernández, A. Millán and A. G. Campaña, Chem. Commun., 2018, 54, 6705–6718.
23. J. Spengler, Y. Wagenhäuser, M. Kuhstrebe and F. Würthner, J. Am. Chem. Soc., 2026, 148, 9014–9022.
24. A. Borissov, Y. K. Maurya, L. Moshniaha, W.-S. Wong, M. Żyła-Karwowska and M. Stępień, Chem. Rev., 2022, 122, 565–788.
25. S. Madayanad Suresh, D. Hall, D. Beljonne, Y. Olivier and E. Zysman-Colman, Adv. Funct. Mater., 2020, 30, 1908677.
26. X.-Y. Wang, X. Yao, A. Narita and K. Müllen, Acc. Chem. Res., 2019, 52, 2491–2505.
27. A. Nowak-Król, P. T. Geppert and K. R. Naveen, Chem. Sci., 2024, 15, 7408–7440.
28. M. Zhao and Q. Miao, Angew. Chem., Int. Ed., 2021, 60, 21289–21294.
29. Y. Yu, A. Izu, J. M. Marín Beloqui, S. Rana, K. S. Mali, S. De Feyter, D. Casanova, J. Casado and J. Liu, J. Am. Chem. Soc., 2025, 147, 20098–20108.
30. X.-Y. Wang, A. Narita, W. Zhang, X. Feng and K. Müllen, J. Am. Chem. Soc., 2016, 138, 9021–9024.
31. C. Chen, J. Lu, Y. Lv, Y. Yan, Q. Sun, A. Narita, K. Müllen and X.-Y. Wang, Angew. Chem., Int. Ed., 2022, 61, e202212594.
32. X. Chen, D. Tan, J. Dong, T. Ma, Y. Duan and D.-T. Yang, J. Phys. Chem. Lett., 2022, 13, 10085–10091.
33. A. S. Scholz, M. Bolte, A. Virovets, E. Peresypkina, H.-W. Lerner, C. S. Anstöter and M. Wagner, J. Am. Chem. Soc., 2024, 146, 12100–12112.
34. A. Kinikar, X.-Y. Wang, M. Di Giovannantonio, J. I. Urgel, P. Liu, K. Eimre, C. A. Pignedoli, S. Stolz, M. Bommert, S. Mishra, Q. Sun, R. Widmer, Z. Qiu, A. Narita, K. Müllen, P. Ruffieux and R. Fasel, ACS Nanosci. Au, 2024, 4, 128–135.
35. H. Lv, K. Xiang and D.-T. Yang, Eur. J. Inorg. Chem., 2022, e202201208.
36. H. Luo and J. Liu, Angew. Chem., Int. Ed., 2024, 63, e202410759.
37. L. Zhang, Y. Cao, N. S. Colella, Y. Liang, J.-L. Brédas, K. N. Houk and A. L. Briseno, Acc. Chem. Res., 2015, 48, 500–509.
38. S. Ito, M. Wehmeier, J. D. Brand, C. Kübel, R. Epsch, J. P. Rabe and K. Müllen, Chem. Eur. J., 2000, 6, 4327–4342.
39. Z. Zhang, D. Csókás, I. Fernández and M. C. Stuparu, Chem, 2024, 10, 3199–3211.
40. Y. Gao, Z. Liu, T. Li and W. Zhao, Angew. Chem., Int. Ed., 2023, 62, e202314006.
41. Z.-L. Qiu, Y. Cheng, Q. Zeng, Q. Wu, X.-J. Zhao, R.-J. Xie, L. Feng, K. Liu and Y.-Z. Tan, J. Am. Chem. Soc., 2023, 145, 3289–3293.
42. M. A. Niyas, K. Shoyama, M. Grüne and F. Würthner, Nature, 2025, 637, 854–859.
43. S. Soldner, K. Shoyama, M. Stolte and F. Würthner, Angew. Chem., Int. Ed., 2026, 65, e2434302.
44. K. Ikemoto, M. Akiyoshi, T. Mio, K. Nishioka, S. Sato and H. Isobe, Angew. Chem., Int. Ed., 2022, 61, e202204035.
45. K. Kato, K. Takaba, S. Maki-Yonekura, N. Mitoma, Y. Nakanishi, T. Nishihara, T. Hatakeyama, T. Kawada, Y. Hijikata, J. Pirillo, L. T. Scott, K. Yonekura, Y. Segawa and K. Itami, J. Am. Chem. Soc., 2021, 143, 5465–5469.
46. K. Zhang, P. A. Hope, M. El Bitar Nehme, A. Linden, B. Spingler and M. Rickhaus, Chem. Eur. J., 2023, 29, e202203954.
47. S. Osumi, S. Saito, C. Dou, K. Matsuo, K. Kume, H. Yoshikawa, K. Awaga and S. Yamaguchi, Chem. Sci., 2016, 7, 219–227.
48. Y. Min, C. Dou, D. Liu, H. Dong and J. Liu, J. Am. Chem. Soc., 2019, 141, 17015–17021.
49. C. Maeda, S. Michishita, I. Yasutomo and T. Ema, Angew. Chem., Int. Ed., 2025, 64, e202418546.
50. M. N. Eliseeva and L. T. Scott, J. Am. Chem. Soc., 2012, 134, 15169–15172.
51. S. Ma, Y. Zhu, W. Dou, Y. Chen, X. He and J. Wang, Sci. China Chem., 2021, 64, 576–580.
52. B. Pigulski, K. Shoyama, M.-J. Sun and F. Würthner, J. Am. Chem. Soc., 2022, 144, 5718–5722.
53. M. Mahl, M. A. Niyas, K. Shoyama and F. Würthner, Nat. Chem., 2022, 14, 457–462.
54. J. C. Hanson and C. E. Nordman, Acta Crystallogr., Sect. B, 1976, 32, 1147.
55. M. A. Petrukhina, K. W. Andreini, J. Mack and L. T. Scott, J. Org. Chem., 2005, 70, 5713–5716.
56. Q. Zhang, K. Kawasumi, Y. Segawa, K. Itami and L. T. Scott, J. Am. Chem. Soc., 2012, 134, 15664–15667.
57. E. Solel, D. Pappo, O. Reany, T. Mejuch, R. Gershoni-Poranne, M. Botoshansky, A. Stanger and E. Keinan, Chem. Sci., 2020, 11, 13015–13025.
58. G. Li, K. Xu, J. Zheng, X. Fang, W. Lou, F. Zhan, C. Deng, Y.-F. Yang, Q. Zhang and Y. She, J. Am. Chem. Soc., 2024, 146, 1667–1680.
59. J. Zhao, Q. Yang, W. Chen, N. Xu, Q. Zhang, W. Zhao, G.-G. Luo, Q. Cui and J. Huang, Chem. Sci., 2025, 16, 20445–20456.
60. M. S. Gordon, D. G. Fedorov, S. R. Pruitt and L. V. Slipchenko, Chem. Rev., 2012, 112, 632–672.
61. (a) CCDC 2486275: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2pg5fs; (b) CCDC 2486277: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2pg5hv; (c) CCDC 2486279: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2pg5kx.

---

Footnote

† K. Zhang and Z.-C. Chen contributed equally to this work.

---