Synthesis and crystal structure of Pe[3]. A triangular macrocycle Pe[3] bearing three pyrene units in the skeletal structure was selected as the co-crystal component as the pyrene group functions as an electron-rich donor and luminophore. Modular synthetic strategies previously reported by us37,38 for the fabrication of macrocycles were followed to synthesize Pe[3] (Supplementary Figs. 1 and 2). In the solid-state structure, Pe[3] crystallizes in the trigonal R-3 space group. Two pairs of enantiomers were tightly stacked to form a supramolecular tetramer characterized by a one-dimensional tubular structure developed exploiting C–H···π, C–H···O, and π···π stacking interactions (Supplementary Figs. 3 and 4). The packing of the supramolecular tetramer in the ab plane accounted for the formation of a four-layer honeycomb 2D architecture. The molecular arrangements of each layer are presented in the Supplementary Fig. 5.
Solid-state superstructures of MCCs. We attempted to develop macrocycle-based co-crystals by carefully regulating the crystallization conditions and using pyrene and TCNB units as the building blocks, as these were suitable conformers that could be efficiently used for the development of co-crystals exploiting CT interactions.39 It was found that solvent modulation played a critical role in controlling the D-A stoichiometric ratio and molecular arrangement in the crystal superstructures. Co-crystallization of electron-rich Pe[3] with electron-deficient TCNB at a ratio of 1:3 in tetrahydrofuran (THF), dioxane, and CHCl3 resulted in the formation of three sets of CT co-crystals, represented as MCC-1, MCC-2, and MCC-3. The Pe[3]:TCNB molar ratios in the crystal structures were 2:1, 1:1, and 2:3, respectively (Supplementary Figs. 6–8). The crystallographic data corresponding to the three co-crystals are presented in the Supplementary Tables 1–3.
The MCCs crystallized in the triclinic P-1 space group. One Pe[3] and 0.5 TCNB molecules were associated with the asymmetric unit of MCC-1, two TCNB (occupancy factor: 0.5) were associated with MCC-2, and 1.5 TCNB were associated with MCC-3 (Supplementary Figs. 9–11). One pair of enantiomers of Pe[3] constituted the three MCCs (Supplementary Fig. 12). It was noted that the conformations of Pe[3] in the four crystal structures (conformations of individual Pe[3] and three binary MCCs units) were different from each other. The methylene angle and the dihedral angle between the dimethoxybenzene and pyrene planes were different (Supplementary Fig. 13), proving that the macrocycle was flexible and adaptive.
The remarkable difference between the structures of the three co-crystals can be attributed to the degree of CT realized and the molecular packing modes for Pe[3] and TCNB (Fig. 1). In MCC-1, one TCNB molecule binds with two Pe[3] molecules to form a sandwich-typed CT complex exploiting the face-to-face π···π interactions. The interplanar distance was calculated to be 3.28 Å (dihedral angle: 1.58°, Fig. 1a), and the molecule consisted of two C–H···O interactions. The distance between the hydrogen atoms in the TCNB units and the closest methoxy oxygen atoms in Pe[3] was calculated to be 2.73 Å (Supplementary Fig. 14). The Pe[3] molecule can interact with an adjacent Pe[3] unit to produce a parallelogram consisting of a supramolecular dimer between the pyrene units via the generation of parallel π···π stacking interactions. The average plane–plane distance was found to be 3.40 Å, and the corresponding dihedral angle was 0° (Supplementary Fig. 15). In the case of MCC-2, the two edges of Pe[3] interact with two types of crystallographically distinct TCNB molecules (TCNB-1 and TCNB-2) in a face-to-face fashion by CT interactions (Fig. 1b). Multiple C–H···N interactions between the nitrogen atoms of the cyano groups in the TCNB molecule and the closest methoxy and pyrenyl hydrogen atoms in Pe[3] further stabilized the CT complex (Supplementary Fig. 16). Each Pe[3] molecule interacted with neighboring Pe[3] molecules via C–H···O, C–H···π, and π···π interactions, forming a tetrameric basic unit (Supplementary Fig. 17). In the crystal structure of MCC-3, the three edges of the Pe[3] unit participated in generating CT interactions with two types of crystallographically distinct TCNB molecules (TCNB-3 and TCNB-4). The units interacted face-to-face to form a butterfly-shaped structure (Fig. 1c). Multiple C–H···N and C–H···O interactions between Pe[3] and TCNB stabilized the CT complex (Supplementary Figs. 18 and 19). The propagation of these non-covalent interactions (in the MCC-1, MCC-2, and MCC-3 molecules) in a 2D plane resulted in the formation of three different 2D superstructure networks (Fig. 1e–g).
Solid-state superstructures of Pe-TCNB. To illustrate that the structural diversity of MCCs can be attributed to the topology of the macrocycles, we grew co-crystals of the monomer (Pe, Scheme 1) with TCNB in the above mentioned solvents (THF, dioxane and CHCl3) under the same crystallization conditions. However, after many attempts, only binary CT co-crystals of Pe-TCNB were obtained in dioxane (Supplementary Fig. 20 and Supplementary Table 4). The Pe:TCNB ratio in the Pe-TCNB system was 1:1, and this was similar to the case of the traditional small-molecule CT co-crystals. The Pe and TCNB molecules were arranged alternately to form a column-like mixed stack (D–A–D–A–D) (Supplementary Fig. S21). The stacks were formed through the exploitation of π···π and C–H···N interactions (Fig. 1d). The columnar structure was extended via the formation of C–H···O interactions to form a 2D assembly (Fig. 1h and Supplementary Fig. 21). The results indicate that the polygonal skeleton of the macrocycles dictates the formation of structurally diverse MCCs.
Photophysical properties of MCCs. Properties of CT-induced supramolecular assemblies have been demonstrated to modulate the optical characteristics of molecules.40 Therefore, we explored the luminescence properties of the MCCs. The fluorescence microscopy images revealed that the three as-prepared co-crystals exhibited tunable luminescence properties. MCC-1 exhibited yellow luminescence, MCC-2 exhibited orange luminescence, and MCC-3 exhibited red luminescence (Fig. 2a). Individual Pe[3] crystals exhibited blue luminescence, and the results indicated that co-crystallization significantly affected the optical properties of the molecules. The luminescence behaviors of organic co-crystals are primarily influenced by the properties of either one donor and several acceptors or one acceptor and several donors that participate in CT interactions.41,42 It has been rarely reported that organic co-crystals characterized by identical supramolecular co-constitutions and varying D-A stoichiometries exhibit different luminescence behavior.43
Solid-state ultraviolet-visible (UV-vis) absorption and fluorescence spectra were recorded and analyzed to elucidate the photophysical properties of the molecules before and after the formation of MCCs (Supplementary Figs. 22–27). As shown in the Supplementary Figs. 22–24, the three co-crystals displayed strong absorption bands, and the peaks corresponding to absorption appeared at 465 (for MCC-1), 497 (for MCC-2), and 532 nm (for MCC-3). The peaks were significantly red-shifted (by 159, 191, and 226 nm, respectively) compared to the peaks corresponding to the individual components of Pe[3] crystals. The red shift could be attributed to the intermolecular CT interactions between Pe[3] and TCNB. The differences in the degrees of red-shift observed for the three CT co-crystals can be attributed to the more efficient D-A [π···π] overlaps in the crystal superstructures of MCC-3 than in that of MCC-2 and MCC-1.
Table 1
Luminescent properties of the three MCCs.
Co-crystal | λem (nm) | ΦF (%) | τav (ns) |
MCC-1 | 575 | 1.2 | 7.38 |
MCC-2 | 602 | 2.7 | 9.38 |
MCC-3 | 635 | 1.0 | 12.66 |
Analysis of the fluorescence emission spectral profiles revealed bathochromic shifts of the peaks appearing in the profiles recorded for the three MCCs (Fig. 2b). Emission peaks appeared at 575, 602, and 635 nm, and the peaks corresponding to the individual Pe[3] crystals appeared at 364 nm. The results agreed well with the results obtained by analyzing the fluorescence microscopy images (Fig. 2a). The International Commission on Illumination (CIE) coordinates of the Pe[3] crystal, and the MCC-1, MCC-2, and MCC-3 co-crystals based were found to be (0.16, 0.06), (0.51, 0.49), (0.54, 0.46), and (0.63, 0.37), respectively (Supplementary Fig. 28), based on the fluorescence spectral profiles. The results revealed that the luminescence color of the co-crystals could be tuned. Meanwhile, the solid-state photoluminescence quantum yield (PLQY, ФF) and fluorescence lifetime (τ) were also determined to better understand the optical properties (Supplementary Figs. 29–36). The results are listed in Table 1, and the corresponding decay curves are presented in Fig. 2c. The solid-state PLQYs of MCC-1, MCC-2, and MCC-3 were 1.2, 2.7, and 1.0%, respectively. The average lifetimes of the three co-crystals increased from MCC-1 (7.38 ns) to MCC-2 (9.38 ns) and MCC-3 (12.66 ns), reflecting the gradual enhancement of the degree of CT realized between the donor and acceptor units. The results are consistent with the variations of the UV-vis absorption profiles.
Frontier molecular orbitals calculations. DFT calculations were performed using the hybrid B3LYP/6-31G functional to obtain deeper insights into the interplay between the structure and property of the MCCs. The structures required for the calculations were obtained by the single-crystal X-ray diffraction data. As shown in Fig. 3, the molecular orbital (MO) diagrams suggested that the highest occupied molecular orbitals (HOMOs) of MCC-1, MCC-2, and MCC-3 were concentrated on the electron-rich pyrene moiety. The energy levels were − 5.18, − 5.04, and − 4.69 eV, respectively. These values were comparable to the values recorded for Pe[3] (− 4.79 eV). The lowest unoccupied molecular orbitals (LUMOs) of the three co-crystals were distributed on the electron-deficient TCNB molecule, and the approximate energy levels were − 2.67, − 2.74, and − 2.71 eV for MCC-1, MCC-2, and MCC-3, respectively. The results indicated that CT interactions occurred when the charge was transferred from the HOMOs of Pe[3] to the LUMOs of TCNB. The gradual decrease in energy gaps (from 2.51 eV to 2.30 eV and subsequently to 1.98 eV) demonstrated an increased degree of CT. This resulted in red-shifted emissions. The results agree well with the changes in the UV–vis absorption spectral profiles. The decrease in bandgap can be attributed to the difference in the D-A ratio of the MCCs. The results confirmed that the stoichiometric ratios dictated the structure of the molecules and the photophysical characteristics of the MCCs.