Synthesis of a Möbius carbon nanobelt

The work addresses the long-standing challenge of synthesizing a fully fused, fully conjugated all-sp2 carbon Möbius nanobelt, which has been theoretically proposed but not realized due to high intrinsic strain. Molecular nanocarbon science has progressed from cycloparaphenylenes and carbon nanobelts (CNBs) to more complex topologies, yet introducing a Möbius twist into a double-stranded, fully fused aromatic belt remained elusive. The study aims to devise a rational synthetic strategy to access Möbius carbon nanobelts (MCNBs) by leveraging precursor symmetry and ring-size-dependent strain, thereby expanding the topological toolbox of nanocarbon chemistry and enabling investigation of their structural dynamics, photophysics, and chirality.

The authors situate their work within advances in molecular nanocarbon synthesis, including cycloparaphenylenes (first proposed in the 1930s, with major synthetic breakthroughs since 2008) and the first fully fused carbon nanobelt (2017), which established belt topologies with distinct inner and outer faces. Prior Möbius-type molecules include non-conjugated double-stranded systems (Walba et al., 1982) and single-stranded Möbius-aromatic annulenes (Herges et al., 2003) with subsequent studies by Latos-Grażyński, Osuka, and others. Double-stranded aromatic Möbius molecules generally required saturated or chalcogen linkers to alleviate strain, limiting full conjugation. Theoretical proposals of Möbius cyclacenes and belts date to the 1990s–2000s, but a fully fused, all-sp2 Möbius CNB had not been synthesized. The present work builds on a previously established CNB synthesis via reductive intramolecular homocoupling of cyclic dibromoparaphenylene–cis-ethenylene precursors, leveraging parity (even vs odd repeat units) to program belt vs Möbius topology.

• Strategy: Modify the established CNB synthetic route so that macrocyclic precursors comprising dibromoparaphenylene and cis-ethenylene units yield CNBs when the number of repeat units is even and MCNBs when odd.
• Computational design: DFT (B3LYP/6-31G(d)) was used to evaluate intrinsic strain and the strain introduced in the final C–C bond formation via homodesmotic reactions. The heat of formation for the final bond-forming step (ΔHf‡, reported as ΔHFRE) was compared for (n,n)CNBs vs (n,n)MCNBs across sizes. For example: (6,6)CNB, ΔHFRE = 40.2 kcal mol−1; (7,7)MCNB, ΔHFRE = 121.1 kcal mol−1. Based on the known success threshold (~40 kcal mol−1 for Ni-mediated homocoupling), target sizes selected were (15,15)MCNB (ΔHFRE = 51.1 kcal mol−1; total strain ≈85.7 kcal mol−1) and (25,25)MCNB (ΔHFRE = 29.6 kcal mol−1; total strain ≈49.4 kcal mol−1).
• Synthesis overview: To improve solubility, n-butoxy groups were installed. The target (BuO)20(25,25)MCNB (1) was synthesized in 14 steps from precursors 2 and 5. • Key building block functionalization: Unsymmetric functionalization of phenanthrene 2 to enable Z-selective Wittig couplings.
  • Monoformylation: TiCl4 and MeOCHCl2 in CH2Cl2 at −45 °C afforded 3 in 75% yield.
  • Chloromethylation: ZrCl4 and MeOCH2Cl in 1,2-dichloroethane at rt gave 4a (bifunctional phenanthrene) in 84% yield.
  • Group interconversions: Convert formyl and chloromethyl to acetal and phosphonium to give 4b. Conditions included TsOH/CH2(OMe)2 (THF/MeOH), PPh3/CH2(OMe)2 (80–90 °C), and subsequent steps (see Fig. 3 conditions: DBU, HCl(aq), TMG, TFA as appropriate). • Oligomer assembly by sequential Z-selective Wittig reactions: React 5 with 4a then 4b to afford key intermediate 7c. From 7c (monomer), prepare dimer (8c), trimer (9c), and pentamer (10c) via further Wittig couplings; functional groups (formyl/phosphonium) react selectively while chloromethyl/dimethylacetal remain inert under those conditions. • Macrocyclization: Derive 10d (bearing formyl and phosphonium) from 10c; macrocyclize with Pr2NEt and MS4A in CHCl3 at 0 °C to form macrocycle 11 in 67% yield (based on 10a). • Reductive intramolecular homocoupling: Treat 11 with Ni(cod)2 and 4,4'-methoxycarbonyl-2,2'-bipyridyl in NMP at 70 °C to give (BuO)20(25,25)MCNB (1) in 20% yield. • Attempted smaller MCNB: Macrocycle 12 (for (15,15)MCNB) gave only a trace mass peak under similar conditions, consistent with prohibitive strain (ΔH‡ = 51.1 kcal mol−1 by DFT).
• Characterization and analysis: • High-resolution mass spectrometry: Observed isotope pattern with highest peak at m/z 3,944.9449, matching calculated m/z 3,944.9423 for C280H260O20. • DFT geometry: Optimized structure of 1 (butoxy → methoxy substituted in calculation) shows C2 symmetry with long axis ~38.4 Å and short axis ~30.3 Å. • Variable-temperature 1H NMR (1,1,2,2-tetrachloroethane-d2): Broad aromatic signals at 25 °C coalesce at 140 °C into seven singlets (assignable to protons a–h by DFT), indicating rapid migration of the Möbius twist around the belt at high temperature. • Dynamics simulation: DFTB-MD simulations visualize twist motion over 0–100 ps. • Photophysics: UV–vis absorption maxima at 389 and 409 nm with a small band at 477 nm; fluorescence (λex 380 nm) with maxima at 480, 513, and 551 nm; fluorescence quantum yield ΦF ≈ 10%; lifetime τ ≈ 14.1 ns; derived kr ≈ 7.1 × 10^6 s−1 and knr ≈ 6.4 × 10^6 s−1. TD-DFT indicates the S1 → S0 transition (477 nm band) is symmetry allowed (oscillator strength f = 0.6239) due to lowered symmetry from the Möbius topology. • Chirality: Chiral HPLC resolves enantiomers; circular dichroism (CD) spectra recorded for each fraction; TD-DFT-simulated CD supports tentative assignment of first and second fractions to M and P helicities, respectively.

• Successful synthesis, isolation, and characterization of a fully fused Möbius carbon nanobelt, (25,25)MCNB (1), achieved over 14 steps using a macrocyclic precursor with an odd number of repeat units and Ni-mediated intramolecular homocoupling.
• Computational strain analysis (B3LYP/6-31G(d)) shows MCNBs have higher strain than same-sized CNBs; final bond formation strain (ΔHFRE) examples: (6,6)CNB 40.2 kcal mol−1 vs (7,7)MCNB 121.1 kcal mol−1. Target sizes chosen accordingly: (25,25)MCNB ΔHFRE 29.6 kcal mol−1 (total strain ≈49.4 kcal mol−1) vs (15,15)MCNB ΔHFRE 51.1 kcal mol−1 (total strain ≈85.7 kcal mol−1).
• Synthesis outcomes: Macrocyclization of 10d to 11 in 67% yield; Ni-mediated homocoupling to (BuO)20(25,25)MCNB (1) in 20% yield. Attempted (15,15)MCNB formation from macrocycle 12 unsuccessful (only trace mass peak), consistent with too-high strain.
• Structural/dynamic characterization: HRMS m/z observed 3,944.9449 (calc. 3,944.9423 for C280H260O20). DFT-optimized 1 exhibits C2 symmetry with axes ~38.4 Å and ~30.3 Å. VT 1H NMR shows coalescence to seven singlets at 140 °C, indicating rapid migration of the Möbius twist; DFTB-MD supports twist mobility.
• Photophysics: Absorption maxima at 389, 409, and a small band at 477 nm; emission maxima at 480, 513, 551 nm (λex 380 nm); ΦF ≈ 10%, τ ≈ 14.1 ns; kr ≈ 7.1 × 10^6 s−1, knr ≈ 6.4 × 10^6 s−1. TD-DFT attributes the 477 nm band to an allowed S1 → S0 transition (f = 0.6239) due to lowered symmetry from Möbius topology.
• Chirality: Chiral HPLC separation accomplished; CD spectra confirm topological chirality with fractions tentatively assigned to M and P enantiomers based on TD-DFT.

The study demonstrates that programming the parity of repeat units in fully fused cyclic precursors enables topological control: even counts yield conventional CNBs, whereas odd counts produce Möbius belts. DFT-guided size selection was critical to overcoming synthetic barriers associated with strain in the final C–C bond formation; larger MCNBs (e.g., (25,25)) are within the accessible strain window for Ni-mediated homocoupling, whereas smaller ones (e.g., (15,15)) are not. The resulting (25,25)MCNB exhibits hallmark features of Möbius topology: non-orientability leading to topological chirality (experimentally validated via chiral HPLC and CD) and dynamic migration of the twist moiety around the belt (observed by VT NMR and supported by DFTB-MD). The photophysical behavior reflects symmetry lowering relative to D3-symmetric CNBs, rendering the S1 → S0 transition allowed and producing a distinct absorption feature at ~477 nm. Collectively, these results confirm both the feasibility of constructing fully conjugated all-sp2 Möbius belts and the profound impact of topology on dynamics and electronic structure, opening avenues for topology-driven nanocarbon materials design.

A fully fused, conjugated Möbius carbon nanobelt was synthesized for the first time using an odd-unit macrocyclic precursor strategy adapted from CNB synthesis. Guided by DFT strain analysis, the (25,25)MCNB was realized and shown to possess dynamic twist migration, distinct photophysical signatures due to symmetry lowering, and resolvable topological chirality. This work establishes a rational, parity-based route to complex nanocarbon topologies and suggests that larger, lower-strain MCNBs and related functionalized variants should be accessible. Future research may explore size tuning, heteroatom or functional group incorporation, solid-state packing and material properties, and potential applications leveraging the unique chiral and electronic characteristics of Möbius belts.

• High intrinsic strain severely limits accessible MCNB sizes; smaller targets like (15,15)MCNB could not be formed under the employed conditions (ΔH‡ ≈ 51.1 kcal mol−1), yielding only trace indications.
• The final Ni-mediated intramolecular homocoupling provides modest yield (20%) for (25,25)MCNB, indicating room for optimization of macrocyclization and coupling efficiencies.
• Solubility considerations necessitated multiple n-butoxy substituents, which may influence properties relative to unsubstituted belts; unsubstituted or differently substituted MCNBs were not reported here.
• Structural dynamics and photophysics were characterized in solution; solid-state properties and device-relevant behaviors were not addressed.