English

• Covalent organic frameworks (COFs) have emerged as a prominent and dynamic research filed since their discovery in 2005 by the Yaghi group. These materials are characterized by their crystalline porous structures, formed through the covalent bonding of small molecular building blocks into extended networks. COFs are widely recognized for their exceptional properties, including high specific surface area, low density, and pre-designable pore environment, making them exceptionally promising for applications in gas adsorption, catalysis, fluorescence sensing, and wastewater purification [1-3]. The functionality of COFs is rooted in the arrangement of their constituent monomers or achieved through subsequent modification strategies [4]. However, a notable challenge arises from the necessity of subjecting COFs to an extensive high-temperature self-healing process to attain crystalline structures. This process often poses challenges in preserving the reactivity of monomer functional groups when using pre-modification techniques [5,6]. Alternatively, post-modification methods may lead to reduce COF porosity and lower-than-expected performance due to incomplete modifications [7,8]. Thus, the conundrum surrounding the compatibility of COF crystallinity and functionalization represents a significant obstacle in this field. To address these challenges, substantial attention has been directed toward the synthesis of sub-stoichiometric COFs, providing a pathway to precisely integrate functional groups within their porous structures. Sub-stoichiometric COFs pertain to structures where a portion of the monomer’s bonding sites remains unreacted during synthesis. Within these COFs, these unreacted functional groups are systematically arranged within the pores, resulting in structures that notably deviate from the originally intended design [9,10]. The phenomenon of sub-stoichiometric COFs was initially observed in 2018 in the context of [4 + 4] type COFs, where two formyl groups from a 4-connected building block did not participate in the synthesis, giving rise to COFs akin to [4 + 2] type COFs [11], now categorized as sub-stoichiometric COFs. Subsequently, Sub-stoichiometry has also been observed in the synthesis of [4 + 3] type COFs, leading to the unexpected bex topology [12-14]. It is well-established that [4 + 3] type COFs generally showcase fully interconnected 3D topologies, featuring fjh, fcc, tbo, pto, and mhq-z topologies [15-18].

  Further investigations have unveiled a significant sub-stoichiometric phenomenon in the synthesis of [6 + 4] type 3D COFs. Specifically, the 6-connected hexaformylphenyl benzene emerges as a tetratopic building block due to its multitude of connected sites and spatial site-blocking effects, leading to the production of sub-stoichiometric 3D COFs without interpenetration [19]. In a recent study, Chen et al. demonstrated their ability to precisely control the co-combination of 4-connected N,N’,N”,N”’-(dibenzo[g,p]chrysene-2,7,10,15-tetrayl)tetrakis(1,1-diph-enyl-methanimine) with 4-connected 1,1,2,2-tetrakis[4-formyl-(1,1-biphenyl)]ethane in the assembly of 2D COFs. This controlled process enabled the synthesis of amino sub-stoichiometric [2 + 4] DBC-TPE-[NH₂], stoichiometric [4 + 4] DBC-TPE, and formyl sub-stoichiometric [4 + 2] DBC-TPE-[CHO] by adjusting the solvent compositions and monomer ratios [9]. This study sheds light on the potential for meticulous control over stoichiometry and sub-stoichiometry in the domain of 2D COFs. However, the comprehensive exploration of assembly and applications related to the sub-stoichiometry of [4 + 4] type 2D COFs remains less explored and necessitates thorough examination.

  In this study, we present another example of modulated synthesis involving both stoichiometric and sub-stoichiometric 2D COFs. This modulation is achieved through strategic manipulation of solvent composition and monomer ratio in the [4 + 4] type COF synthesis. This meticulous approach resulted in the synthesis of three distinctive COF variants: sub-stoichiometric COF-DA with a [4 + 2] connectivity pattern, stoichiometric COF-DAPy featuring [4 + 4] connectivity, and sub-stoichiometric COF-Py with a [2 + 4] connectivity pattern. The structural features of these COFs were rigorously confirmed and validated through a comprehensive array of characterization techniques, including powder X-ray diffraction (PXRD), Fourier-transform infrared spectroscopy (FT-IR), and solid-state nuclear magnetic resonance (ssNMR). Furthermore, their pore structures were distinctly observed via transmission electron microscopy (TEM). Furthermore, we conducted a comparative analysis of the resulting 2D COFs for gas adsorption and separation. This analysis revealed that sub-stoichiometric COF-DA, characterized by residual amino groups, exhibits gas adsorption capacities for C₂H₂, C₂H₄, and CO₂ gases comparable to those of stoichiometric COF-DAPy. In contrast, sub-stoichiometric COF-Py, containing residual formyl groups, demonstrates enhanced performance in the separation of C₂H₂/CO₂ and C₂H₂/C₂H₄ compared to COF-DAPy.

  The synthesis of three unique 2D COFs, denoted as COF-DA, COF-DAPy, and COF-Py, was achieved through meticulous control of solvent selection and initial monomer feeding ratios, utilizing the conventional solvent-thermal method (Fig. 1 and Schemes S1-S3 in Supporting information). Characteristic C=N stretching vibrations are observed in the FT-IR spectra of the synthesized materials, appearing at around 1624 cm⁻¹ (Fig. S1 in Supporting information). The presence of imine bonds within these frameworks was further confirmed using solid-state ¹³C CP-MAS NMR spectroscopy, revealing distinctive peaks corresponding to the C=N moieties at chemical shifts of 155, 155, and 157 ppm for COF-DA, COF-DAPy, and COF-Py, respectively (Fig. S2 in Supporting information). Remarkably, COF-Py exhibits prominent peaks associated with formyl groups in both the FT-IR spectrum (1697 cm⁻¹) and the solid-state ¹³C CP/MAS NMR spectrum (190 and 192 ppm), indicating a substantial presence of residual formyl groups within its skeleton. The thermal stability of these synthesized COFs was assessed through thermogravimetric analysis (TGA), revealing their exceptional stability over 500 ℃. It is noteworthy that COF-DA displays mass loss at 170 ℃, attributed to the decomposition of amino groups within its structure, whereas COF-Py exhibits mass loss at 387 ℃ due to formyl groups decomposition (Fig. S3 in Supporting information).

  Figure 1

  

  Figure 1.  Schematic representation of the synthesis of stoichiometric and sub-stoichiometric 2D COFs employing identical building blocks, resulting in sub-stoichiometric [2 + 4] type COF-DA with remaining amino groups, stoichiometric [4 + 4] type COF-DAPy, and sub-stoichiometric [4 + 2] type COF-Py with residual formyl groups.

  DownLoad: Full-Size Img PowerPoint

  The crystal structures and unit cell parameters of COF-DA, COF-DAPy, and COF-Py were determined through PXRD measurements and subsequent structural simulations. The PXRD patterns of COF-DA and COF-Py, while displaying some similarities, exhibit significant deviations from that of COF-DAPy, shedding initial light on the distinct topologies of stoichiometric and sub-stoichiometric COFs (Figs. 2a-c). Considering the symmetries and connectivity of the constituent building blocks, three distinct structural models, namely, sub-sql-α, sql, and sub-sql-β were proposed. In these models, COF-DA and COF-Py were conceived as sub-stoichiometric structures with remaining amino and formyl groups, respectively. Additionally, the AA stacking and AB stacking structures of COF-DAPy were constructed using the sql topology (Figs. S4-S6 in Supporting information), aligning with previous reports [20]. Experimental PXRD patterns of COF-DA and COF-Py closely match the simulated patterns with TPPDA and TFPPy as sub-stoichiometric combinations, respectively. Notably, the primary experimental diffraction peaks observed at 3.28°, 5.53°, 6.53°, 8.61°, 11.75°, 14.84°, 20.86° for COF-DA (red curve in Fig. 2a) and 3.34°, 5.63°, 6.58°, 8.18°, 11.72°, 14.45°, 20.93° for COF-Py (blue curve in Fig. 2c) can be assigned to the (220), (310), (510), (440), (910), and (001) reflection planes, respectively. Further confirmation was derived through Pawley refinement, displaying a remarkable agreement between the experimental and refined PXRD profiles, evidenced by the convergence of R_wp and R_p values to 4.73% and 3.45% for COF-DA, and 5.08% and 3.92% for COF-Py, respectively. The refined unit cells parameters, determined using the Pawley method, are presented in Tables S1 and S3 (Supporting information). Specifically, for COF-DA, the parameters are a = 53.1338 Å, b = 33.7352 Å, c = 3.7385 Å, while for COF-Py, they are a = 55.3523 Å, b = 37.0328 Å, c = 3.5945 Å, showing both in C222 (No. 21) space group. A comparable method was employed to establish the sql topology of COF-DAPy with an AA stacking mode, yielding negligible divergences with R = 3.65% and R = 2.87%. Refinement results indicate that COF-DAPy possesses unit cell parameters of a = 26.3952 Å, b = 21.6966 Å, c = 3.6516 Å, and pertains to the P222 (No. 16) space group (Table S2 in Supporting information).

  Figure 2

  

  Figure 2.  (a-c) Experimental and simulated PXRD patterns of (a) COF-DA, (b) COF-DAPy, and (c) COF-Py. (d-f) N₂ adsorption-desorption isotherms measured at 77 K and the pore size distributions of (d) COF-DA, (e) COF-DAPy, and (f) COF-Py. (g-i) Structural models and pore aperties of (g) COF-DA, (h) COF-DAPy, and (i) COF-Py.

  DownLoad: Full-Size Img PowerPoint

  The structural models ascribed to COF-DA, COF-DAPy, and COF-Py were further substantiated through the examination of N₂ sorption isotherms at 77 K. All three COFs exhibit reversible type Ⅰ/Ⅳ sorption isotherms, signifying their microporous or mesoporous nature. The Brunauer-Emmet-Teller (BET) specific surface areas were calculated to be 752 m²/g, 984 m²/g, and 1106 m²/g for COF-DA, COF-DAPy, and COF-Py, respectively. The pore size distribution, assessed using non-local density functional theory (NLDFT), reveals two primary pore sizes centered at 1.0 nm and 1.9 nm for COF-DA and COF-Py, and a single prominent pore size centered at 1.5 nm for COF-DAPy (Figs. 2d-f). These findings closely aligned with the theoretical pore sizes derived from their respective modeled structures (Figs. 2g-i). Scanning electron microscopy (SEM) analyses were employed to investigate the morphologies of these three COFs (Figs. 3a-c and Fig. S7 in Supporting information). Evidently, all three COFs reveal a rod-like aggregated morphology composed of small lamellar crystals. Notably, COF-DAPy presents a distinct lycopod-like morphology, setting it apart from COF-DA and COF-Py, which display a relatively flaky structure.

  Figure 3

  

  Figure 3.  (a-c) SEM images of (a) COF-DA, (b) COF-DAPy, and (c) COF-Py. (d-f) TEM image of (d) COF-DA, (e) COF-DAPy, and (f) COF-Py. Inset: the corresponding FFT patterns. (g-i) Enlarged TEM images of selected areas in d-f with theoretical structural models inseted.

  DownLoad: Full-Size Img PowerPoint

  This discrepancy in morphology can be ascribed to the stronger π-π stacking and more uniform inter-crystal forces in the stoichiometric COF-DAPy. In-depth microstructural analysis of these three COFs was conducted using transmission electron microscopy (Figs. 3d-i and Fig. S8 in Supporting information). Thanks to the remarkably high crystallinity of these synthesized COFs, TEM facilitated the detailed observation of fine structural information and provided insights into the structural characteristics of these COFs on the (001) crystal plane (Figs. 3d-f). Remarkably, the pore information observed in magnified regions closely aligns with the theoretically modeled structures of these two COFs (Figs. 3g and h). Furthermore, the corresponding fast Fourier transform (FFT) images exhibit precise alignment with the simulated electron diffraction patterns (Fig. S9 in Supporting information), further confirming the accuracy of the structural models. Similarly, Figs. 3e and i highlight the remarkable crystallinity of COF-DAPy, revealing a rhombic pore structure that correlates well with the stoichiometric sql topology.

  Based on the abovementioned analyses, it is strongly substantiated that two distinct sub-stoichiometric COFs have been effectively constructed utilizing identical building blocks. Despite the resemblance in their structures and PXRD patterns, COF-DA and COF-Py reveal diverse photoluminescent properties, primarily arising from the distinct configurations of the pyrene chromophoric unit within the COF frameworks. The constrained pyrene conformations in COF-DA and COF-DAPy, in their aggregated state, hamper photoluminescence due to robust interlayer π-π stacking and the aggregation-caused quenching (ACQ) effect. In contrast, COF-Py, endowed with a greater degree of freedom in pyrene motifs, exhibits a prominent aggregation-induced emission (AIE) effect [21], and thus emitting vivid yellow fluorescence under UV irradiation at 375 nm (Fig. S10 in Supporting information). The optical characteristics of these three COFs were further assessed using ultraviolet-visible absorption spectroscopy (UV–vis) and X-ray photoelectron spectroscopy (XPS) techniques. Analysis of UV–vis diffuse reflectance spectra reveals distinctive light-absorption profiles for COF-DA, COF-DAPy, and COF-Py. The optical band gaps, determined via the Kubelka-Munk function, are calculated to be 1.93 eV, 2.11 eV, and 1.85 eV, respectively (Figs. S11a and b in Supporting information). The detailed band edge positions for these COFs were further deduced via the VB-XPS analysis coupled with the UV–vis results (Figs. S11c and d in Supporting information). The XPS spectra validate that these COFs exclusively consist of carbon (C), nitrogen (N), and oxygen (O) elements, indicative of their high purity. Notably, the detailed C 1s and N 1s spectra clearly demonstrate the presence of C=N bonds, underscoring the successful synthesis of COFs. Furthermore, the N 1s XPS spectrum of COF-DA unveils the existence of amino groups at 400 eV, a feature not observed in the spectra of COF-DAPy and COF-Py (Figs. S12-S14 in Supporting information). These findings provide robust support for our conclusion that precise manipulation of the monomer ratios enables the synthesis of three distinct 2D COFs.

  The intrinsic porosity and the abundance of amino/formyl groups residing within COF pores prompted our exploration of their potential application in gas adsorption and separation at low pressure. A systematic evaluation of the adsorption behaviors of these COFs concerning C₂H₂, C₂H₄, and CO₂ gases was conducted at 298 K. All three COFs exhibit higher C₂H₂ uptake capacities than those of C₂H₄ and CO₂, with the adsorption order being C₂H₂ > C₂H₄ > CO₂. This trend can be ascribed to the increased polarizability and substantial quadruple moment of C₂H₂ (Figs. 4a-c). The adsorption capacities for C₂H₂, C₂H₄, and CO₂ were quantified as 56.6, 46.5, and 24.2 cm³/g for COF-DA, 54.1, 47.4, and 25.8 cm³/g for COF-DAPy, and 45.2, 24.6, and 18.3 cm³/g for COF-Py at the ambient conditions. It is noteworthy that COF-Py exhibits a considerably distinct adsorption capacity for C₂H₂ in contrast to C₂H₄ when compared to COF-DA and COF-DAPy, which may be attributed to the abundent accessible formyl functional groups within COF-Py. To assess the feasibility of above separation, ideal adsorbed solution theory (IAST) calculations were employed to predict the selectivity of equimolar binary gas mixtures of C₂H₂/C₂H₄ and C₂H₂/CO₂ at 298 K (Figs. 4d-f and Figs. S15-S17 in Supporting information). Due to the greater accessibility of residual formyl groups in COF-Py, significantly improved selectivities were observed. Consequently, at 1 atm and 298 K, COF-DA and COF-DAPy exhibited C₂H₂/C₂H₄ and C₂H₂/CO₂ selectivities of 2.4 and 1.1, and 2.6 and 1.1, respectively, while COF-Py exhibited notably higher selectivities, with C₂H₂/C₂H₄ and C₂H₂/CO₂ selectivities of 2.9 and 2.8, respectively (Figs. 4d-f). The C₂H₂/C₂H₄ selectivity of COF-Py ranks among the highest compared to many reported COF materials like NKCOF-62 (1.3) [22], NPU-1 (1.4) [23], and UPC-613 (1.5) [24], underlining the potential of COF-Py for applications in ethylene purification.

  Figure 4

  

  Figure 4.  (a-c) C₂H₂, C₂H_4, and, CO₂ sorption profiles for (a) COF-DA, (b) COF-DAPy, and (c) COF-Py measured at 298 K. (d-f) Selectivities of equimolar binary C₂H₂/C₂H₄ and C₂H₂/CO₂ for (d) COF-DA, (e) COF-DAPy, and (f) COF-Py at 298 K.

  DownLoad: Full-Size Img PowerPoint

  In summary, our study has successfully realized the precise synthesis of stoichiometric and two sub-stoichiometric 2D COFs through meticulous adjustments of solvent compositions and monomer ratios in the [4 + 4] condensation reaction. It is noteworthy that, in comparison to stoichiometric COF-DAPy, sub-stoichiometric COF-DA, characterized by residual amino groups, demonstrates increased gas adsorption capacities, while sub-stoichiometric COF-Py, featuring unreacted formyl groups, exhibits enhanced adsorption selectivity in the separation of C₂H₂ from C₂H₄. This study not only provides new examples for customizing the synthesis of both stoichiometric and sub-stoichiometric 2D COFs but also promotes the exploration of these tailored COF materials in the efficient purification of ethylene.

  Declaration of competing interest

  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.

  Acknowledgments

  This work was supported by the National Key Research and Development Project of China (No. 2022YFE0113800), the National Natural Science Foundation of China (No. 22375179), and the start-up grant (No. 2019125016829) in Zhejiang University of Technology. This work is partially supported by the National Innovation and Entrepreneurship Training Program (No. 202310337063).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109760.

  
  
• 1. [1]

     Q. Du, J. Zou, Z. Huang, et al., Chin. Chem. Lett. 34 (2023) 107763. doi: 10.1016/j.cclet.2022.107763

  2. [2]

     J. Li, Y. He, Y. Zou, et al., Chin. Chem. Lett. 33 (2022) 3017–3020. doi: 10.1016/j.cclet.2021.12.011

  3. [3]

     P. Shang, X. Yan, Y. Li, et al., Chin. Chem. Lett. 34 (2023) 107584. doi: 10.1016/j.cclet.2022.06.007

  4. [4]

     W.R. Cui, C.R. Zhang, W. Jiang, et al., Nat. Commun. 11 (2020) 436. doi: 10.1038/s41467-020-14289-x

  5. [5]

     L. Deng, Z. Ding, X. Ye, et al., Acc. Mater. Res. 3 (2022) 879–893. doi: 10.1021/accountsmr.2c00108

  6. [6]

     H. Vardhan, A. Nafady, A.M. Al-Enizi, et al., Nanoscale 11 (2019) 21679–21708. doi: 10.1039/c9nr07525a

  7. [7]

     H. Lyu, H. Li, N. Hanikel, et al., J. Am. Chem. Soc. 144 (2022) 12989–12995. doi: 10.1021/jacs.2c05382

  8. [8]

     D. Dutta, J. Wang, X. Li, et al., Small 18 (2022) e2202369. doi: 10.1002/smll.202202369

  9. [9]

     X. Yang, Z. Xie, T. Zhang, et al., Sci. China Chem. 65 (2022) 190–196. doi: 10.1007/s11426-021-1064-y

  10. [10]

      M. Wu, Z. Shan, J. Wang, et al., Chem. Eng. J. 454 (2023) 140121. doi: 10.1016/j.cej.2022.140121

  11. [11]

      Q. Gao, X. Li, G.H. Ning, et al., Chem. Mater. 30 (2018) 1762–1768. doi: 10.1021/acs.chemmater.8b00117

  12. [12]

      Y. Liu, W.K. Han, W. Chi, et al., Appl. Catal. B 331 (2023) 122691. doi: 10.1016/j.apcatb.2023.122691

  13. [13]

      M. Lu, M. Zhang, J. Liu, et al., J. Am. Chem. Soc. 144 (2022) 1861–1871. doi: 10.1021/jacs.1c11987

  14. [14]

      T. Banerjee, F. Haase, S. Trenker, et al., Nat. Commun. 10 (2019) 2689. doi: 10.1038/s41467-019-10574-6

  15. [15]

      H.L. Nguyen, C. Gropp, O.M. Yaghi, J. Am. Chem. Soc. 142 (2020) 2771–2776. doi: 10.1021/jacs.9b13971

  16. [16]

      J. Chao, Z.K. Wang, H. Liu, et al., J. Catal. 413 (2022) 692–702. doi: 10.1016/j.jcat.2022.07.023

  17. [17]

      Q. Liao, W. Xu, X. Huang, et al., Sci. China Chem. 63 (2020) 707–714. doi: 10.1007/s11426-019-9696-3

  18. [18]

      D. Zhu, Y. Zhu, Y. Chen, et al., Nat. Commun. 14 (2023) 2865. doi: 10.1038/s41467-023-38538-x

  19. [19]

      L. Chen, C. Gong, X. Wang, et al., J. Am. Chem. Soc. 143 (2021) 10243–10249. doi: 10.1021/jacs.1c03739

  20. [20]

      A.F.M. El-Mahdy, M.G. Mohamed, T.H. Mansoure, et al., Chem. Commun. 55 (2019) 14890–14893. doi: 10.1039/c9cc08107k

  21. [21]

      Y. Wang, Y.Z. Cheng, K.M. Wu, et al., Angew. Chem. Int. Ed. 62 (2023) e202310794. doi: 10.1002/anie.202310794

  22. [22]

      Z. Wang, Y. Zhang, E. Lin, et al., J. Am. Chem. Soc. 145 (2023) 21483–21490. doi: 10.1021/jacs.3c07224

  23. [23]

      B. Zhu, J. Cao, W. Mukherjee, et al., J. Am. Chem. Soc. 143 (2021) 1485–1492. doi: 10.1021/jacs.0c11247

  24. [24]

      Y. Wang, C. Hao, W. Fan, et al., Angew. Chem. Int. Ed. 133 (2021) 11451–11459. doi: 10.1002/ange.202100782

• 

  Figure 1  Schematic representation of the synthesis of stoichiometric and sub-stoichiometric 2D COFs employing identical building blocks, resulting in sub-stoichiometric [2 + 4] type COF-DA with remaining amino groups, stoichiometric [4 + 4] type COF-DAPy, and sub-stoichiometric [4 + 2] type COF-Py with residual formyl groups.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a-c) Experimental and simulated PXRD patterns of (a) COF-DA, (b) COF-DAPy, and (c) COF-Py. (d-f) N₂ adsorption-desorption isotherms measured at 77 K and the pore size distributions of (d) COF-DA, (e) COF-DAPy, and (f) COF-Py. (g-i) Structural models and pore aperties of (g) COF-DA, (h) COF-DAPy, and (i) COF-Py.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a-c) SEM images of (a) COF-DA, (b) COF-DAPy, and (c) COF-Py. (d-f) TEM image of (d) COF-DA, (e) COF-DAPy, and (f) COF-Py. Inset: the corresponding FFT patterns. (g-i) Enlarged TEM images of selected areas in d-f with theoretical structural models inseted.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a-c) C₂H₂, C₂H_4, and, CO₂ sorption profiles for (a) COF-DA, (b) COF-DAPy, and (c) COF-Py measured at 298 K. (d-f) Selectivities of equimolar binary C₂H₂/C₂H₄ and C₂H₂/CO₂ for (d) COF-DA, (e) COF-DAPy, and (f) COF-Py at 298 K.

  下载: 全尺寸图片 幻灯片
•