The glamour of dynamic covalent chemistry (DCC) [1–4] comes from its ability of "error-checking" and "self-correction" to produce the most thermodynamically stable product as the major or sometimes only one product out of a library of many other combinations with amazing selectivity and high efficiency. It combines the robustness of covalent bonds in traditional organic chemistry and reversibility of noncovalent interactions popular in supramolecular chemistry. A lot of reversible covalent reactions with rather strong bonding, such as dynamic nucleophilic aromatic substitution (S_NAr) [5,6], boronic ester formation [7–9], imine condensation between amine and carbaldehyde [10], disulfide bridges [11], laid the basis for DCC. Among which, the imine condensation reaction, of course, is the most prominent and widely used one. The labile nature of imines in the presence of water could be overcome by chelating effect of multivalence and ligand preorganization [12,13] and mutual stabilization via coordination with unstable Cu⁺ [14], which opened an even wider door for its applications. It has become a very powerful tool for constructing various structures with amazing complexity such as organic cages [15,16], covalent organic frameworks (COF) [17–19], macrocycles [20], superphanes [21], with high precision and high efficiency. Here we reported one-pot synthesis and structure of a unique twisted double-layer chiral macrocycle from the acid catalyzed [4 + 8] imine condensation between V-shaped tetraaminooxacalix[4]arene and N-alkylcarbazole-3, 6-dicarbaldehyde.

Oxacalixarenes [22–26], a new type of macrocycles by replacing the bridging −CH₂− or −CHR− in classic calixarenes with −O−, have prevailed in supramolecular chemistry since their synthesis breakthrough was achieved at the early of the 21^st century [27,28]. Previously we reported facile and efficient synthesis of a nitro and ester functionalized oxacalix[4]arene [29]. The V-shaped 1,3-alternate conformation and easy derivation make it a good platform for anion and cation sensors [30,31]. We further demonstrated that this easily affordable module was an excellent building block for the construction of robust organic cages via multistep organic syntheses [32]. Here, we wished to synthesize a new type of shape-persistent molecular cage as depicted in Scheme 1 with this building block via dynamic [2 + 4] imine condensation reaction [33,34], which would greatly reduce the synthetic endeavor.

By mixing tetraaminooxacalix[4]arene 1 and N-alkylcarbazole-3,6-dicarbaldehyde 2 in a ratio of 1:2 in various solvents including CHCl₃, THF, CH₃OH, C₂H₅OH, pyridine, toluene, DMSO, CH₃CN, 1,4-dioxane, etc., with or without a catalyst (Table S1 in Supporting information), thin layer chromatography (TLC) monitoring indicated either no reaction or inseparable mixture or polymer formation. When we came to dry CH₂Cl₂ with CF₃COOH (1 mmol/L) as a catalyst, after refluxing in a sealed tube for 48 h under a nitrogen atmosphere, we got almost only one product as indicated by TLC, though there might be many other combinations. After purification via column chromatography, light yellow powders were obtained in 89% and 87% yield for N-(2-ethylhexyl) and N-isobutyl products, respectively. Much to our surprise, ¹H NMR measurements revealed well-defined yet undissolvable spectra for the two products (Fig. 1, middle): though with sharp signals, there are so many sets of peaks that could not be interpretated by the structure of [2 + 4] molecular cage 3 as depicted in Scheme 1 via imine condensation reaction. The products were very sensitive to trace amount of acid in the untreated CDCl₃. The vanishment of CHO signal at δ 10.14 and NH₂ signal at δ 3.88 might indicate completion of imine condensation reaction. Four sets of singlet peaks (labelled with asterisks in Fig. 1) were observed both at δ 8.5~9 region and δ 5.5~6 region, which might come from H^a and H^d in the oxacalix[4]arene skeleton. DOSY measurement on 4b (Fig. S11 in Supporting information) revealed that almost all signals have the same diffusion coefficient of D = (1.29 ± 0.03) × 10⁻¹⁰ m²/s, suggesting only one species in solution. A hydrodynamic radius of 43 Å was calculated from Stokes–Einstein equation [35], indicating formation of a giant discrete structure. With lowering of temperature from 293 K to 203 K in CD₂Cl₂ (Fig. S12 in Supporting information), ¹H NMR spectra of 4b showed only little up field shifting and signal broadening. No more new peaks appeared, which precluded fast equilibrium between various structures. ESI HRMS measurements displayed two signals at m/z = 5711.8999 and 5262.3193 for the two products respectively, which corresponded to the ([4 + 8] + H)⁺ signals. This phenomenon suggests formation of well-defined, discrete, and very complex structures in solution. We tried to introduce a second transformation from imine to amine with NaBH₄, or NaBH₃CN, or catalytic hydrogenation (Pd/C/H₂) to overcome the reversibility of the C=N double bonds. Unfortunately, we got a complex mixture each time.

After many trials, a block-shaped single crystal suitable for X-ray analysis was obtained via slow evaporation of a CH₂Cl₂/CH₃CN solution. It displayed a giant twisted double-layer [4 + 8] macrocycle structure 4b (Fig. 2 and Table S2 in Supporting information, CCDC: 2291973), with a shape of figure eight. The macrocycle is chiral and a pair of enantiomers are found in each unit cell. The four oxacalix[4]arene motifs in each macrocycle are not completely equivalent to each other, which might account for the complicated ¹H NMR spectra discussed above. They show the following common features: they all adopt 1,3-alternate conformations as found in most oxacalix[4]arene skeletons; two ester-bearing aromatic rings are almost antiparallel with dihedral angles in the range of 152.39°–157.98° and centroid−centroid distances in the range of 7.46–7.51 Å; the two aromatic rings bearing imine groups are almost parallel with dihedral angles in the range of 10.50°–14.76° and centroid-centroid distances in the range of 4.49–4.62 Å; the carbazole motifs appear in nearly parallel pairs with a very little dihedral angle of about 2.5° and interlayer distance of about 3.5 Å, a distance for typical π-π stacking interactions.

DFT calculations further revealed energy differences between the reactants and the products, which might shed light on the selective formation of the unique twisted structures. The optimization and frequency analysis for the two reactants, tetraaminooxacalix[4]arene (methyl ester derivative used) and N-methylcarbazole-3,6-dicarbaldehyde, the hypothetic [2 + 4], [3 + 6], [4 + 8] products, the real [4 + 8] twisted product from single crystal, and water were obtained through DFT calculations via B3LYP method at the 6–31G(d, p) level with EmpiricalDispersion = GD3BJ. The single point energies for the above optimized structures were calculated using a higher level of basis set 6–311G(2d, p). The results are summarized in Figs. S14 and S15 and Table S3 (Supporting information). After optimization, the hypothetic [2 + 4] and [3 + 6] macrocycles are almost planar, and the hypothetic [4 + 8] product is a little twisted. The changes in Gibbs free energy and enthalpy for the hypothetic processes are all positive. However, those for the twisted [4 + 8] macrocycle are negative, ∆_rG_m = −76.8 kJ/mol and ∆_rH_m = −139.8 kJ/mol. The striking energy differences for the hypothetic and real processes might account for the amazing high selective and efficient formation of the twisted macrocycle product.

In summary, we reported here a new type of giant twisted double-layer chiral macrocycles, which was formed via one-pot acid catalyzed [4 + 8] condensation reaction between a V-shaped tetraaminooxacalix[4]arene and N-alkylcarbazole-3,6-dicarbaldehyde. Sixteen imine bonds formed in a single step with amazing efficiency as high as 89%. Well-defined yet very complicated ¹H NMR spectra could not be interpretated. X-ray single analysis provided high-resolution structural data for the structure unambiguously. DFT calculations further revealed the thermodynamical origin of the unique high selectivity and efficiency. Subsequent reduction to circumvent the reversibility of the imine bonds failed. The subtle change in the geometry accompanying the reduction disrupted the perfect match and led to complete collapse of the unique structure.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work is supported by the National Natural Science Foundation of China (Nos. 21971223 and 21772178). We thank Dr. Chengshuo Shen for help in DFT calculations.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109575.