Introduction

In recent years, the global climate change mainly caused by excessive carbon dioxide (CO2) emissions has drawn great attentions and concerns1. Developing viable CO2 capture and storage (CCS) technologies to stabilize atmospheric CO2 levels and cope with global warming, which is an effective way2. Porous organic polymers (POPs) constructed by low mass density, non-metallic elements, not only have a large specific surface area, high pore volume, narrow pore size distribution, good chemical and physical stability, and wide synthetic diversification, but also present cost and effective gas uptake applications3,4,5. POPs physisorb CO2 molecules via weak van der Waals forces, which are potential candidates for CO2 capture because of their low regeneration energy consumption and high CO2 sorption capacity6. In the past few years, versatile POPs materials such as covalent organic frameworks (COFs)7,8, covalent triazine-based frameworks (CTFs)6, polymers of intrinsic microporosity (PIMs)9,10, porous aromatic frameworks (PAFs)11, conjugated microporous polymers (CMPs)12, and hypercross-linked polymers (HCPs)13, have been rapidly developed due to their important applications in a broad variety of aspects including gas storage/separations14,15, chemosensors16,17,18, tunable photoluminescence19,20, heterogeneous catalysis21,22 and so on.

Conjugated microporous polymers (CMPs) are a new class of porous materials, which are synthesized by transition metal coupling chemistry including Pd-catalyzed Suzuki and Sonogashira cross-coupling polycondensation23, Ni-catalyzed Yamamoto reaction24, and other reactions such as oxidative polymerization25, Schiff-base reaction26. The unmatched feature of CMPs is that they combine π-conjugation and permanent porous structure in a bulk material. Recently, some reports have revealed that the introduction of some polar functional groups or heteroatoms into porous materials could enhance the binding affinity between the adsorbent and CO2 molecules, which leads to the increase of CO2 capture capacity27,28,29,30. In this work, we chose carbazole as the network core and different flexible monomers as linker space to construct the final network based on the following reasons: (1) carbazole-based porous organic polymers have been studied as strong candidates for carbon dioxide (CO2) due to the rigid structure and special intrinsic properties of their building blocks; (2) carbazole possesses polar group (-NH) from carbazole unit, which might promote the interaction between the solid adsorbent and acidic CO2 molecules; (3) conjugated property of rigid carbazole unit is beneficial for formation of porous structure with permanent porosity and stability; (4) the linker monomers possess different steric configuration, such as linear type, triangle, and tetrahedral, which can exhibit different flexibility to form the microporous volume. Therefore, introduction of the carbazole unit into the polymer skeleton makes the porous materials electron-rich to enhance CO2 uptake of porous polymers, and the steric configuration of linker monomer can effectively construct and adjust the microporous volume and reach outstanding gas storage and separation capacity31,32,33.

With these considerations in mind, herein, we prepared a series of CMPs (Fig. 1, N4CMP-1–5) based on 1,3,6,8-tetrabromocarbazole as the basic buliding block via Pd-catalyzed Suzuki cross-coupling or Sonogashira polycondensation. The porosity properties and surface area in the kind of amorphous CMPs can be finely controlled by the strut length, size or geometry of linker. The obtained carbazole-based CMPs possess large specific surface area, high micropore volume, narrow pore size distribution, high gas uptake capacity, and good selectivity toward CO2 over N2 or CH4. Meanwhile, N4CMP polymer networks also exhibit high CO2 adsorption capacity at high pressure condition.

Figure 1
figure 1

Schematic representation of synthesis of N4CMP polymers.

Full size image

Results

Synthesis and characterization

All of the polymer networks were synthesized by palladium (0)-catalyzed cross-coupling polycondensation of 1,3,6,8-tetrabromocarbazole and a number of benzeneboronic monomers or ethynyl monomers. All the reactions were carried out at a fixed reaction temperature and reaction time (120 °C/48 h). The general synthetic route towards N4CMPs polymers is shown in Scheme 1. Our aim is to explore the effect of structure and connecting position of linker on pore properties of the resulting porous polymers. The insoluble polymers were filtered and washed with water, tetrahydrofunan, chloroform and methanol, respectively, in order to remove the inorganic salts, organic monomers, residual catalyst, and oligomers. All of these polymers are insoluble in common organic solvents because of their highly cross linked structures.

The structures of N4CMPs polymers were further characterized by Fourier transform Infrared (FT-IR) spectroscopy (Fig. 2). All the polymer networks show the characteristic C=C stretching band at 1600 cm−1 and aromatic C–H stretching frequencies up to 3000 cm−1. All the spectra exhibit intense characteristic bands of N-H at about 3450 and 1390 cm−1 31,32,33. The band around 1500 cm−1 is assigned to the stretching vibration of C-N-C in the five-membered NC4 ring for all the samples31,32,33. The primary bromo group of 1,3,6,8-tetrabromocarbazole at about 590 cm−1 are absent in the polymer networks. The typical C≡C stretching mode at about 2200 cm−1 is also observed in the N4CMP-2, N4CMP-3, N4CMP-4, and N4CMP-5, respectively. These results suggested that a hyper-cross-linked structure was successfully obtained. The structures of N4CMPs polymers were characterized at the molecular level by solid state13C cross-polarization magic-angle spinning (CP/MAS) NMR (ESI, Figure S1). In general, N4CMP-1–5 have the similar broad peaks. The resolved resonance peak at about 137 ppm arises from the carbons of the carbazole rings. Two peaks at about 130, 123, and 110 ppm can be ascribed to the other carbons in the aromatic rings. The small peaks at about 93 ppm correspond to acetylene carbons from acetylene monomers. These peaks are perfectly consistent with previous works31,32,33.

Figure 2
figure 2

FT-IR spectra of 1,3,6,8-tetrabromocarbazole (black), and polymers N4CMP-1–5.

Full size image

The morphology information was collected by field-emission scanning electron microscopy (FE-SEM) and high-resolution transmission electron microscopy (HR-TEM). The typical SEM images of N4CMPs indicate that all the polymers have irregular shapes with sizes of about several micrometers (Fig. 3). The TEM images revealed that they are porous structures of the materials, which is similar to some reported amorphous microporous organic materials (ESI, Figure S2)34,35,36. The UV/vis diffuse reflectance spectra of the N4CMP are displayed in Figure S3. Both polymers showed a broad absorption range. All the five polymers displayed a broad absorption with wavelengths ranging from 375 to 428 nm. Although the shapes of the spectra are similar, they do display differences in terms of the onset of the absorption, which correlates with the band gap of the conjugated system. The TGA results verify that the polymers have a good thermal stability, and the thermal degradation temperature is up to ca. 360 °C (ESI, Figure S4). The weight loss below 100 °C is generally attributed to the evaporation of adsorbed water and gas molecules trapped in the micropores. Powder X-ray diffraction (PXRD) measurements indicated that all the polymers are amorphous in nature (ESI, Figure S5), as most other reported CMP networks12.

Figure 3
figure 3

FE-SEM images of (a) N4CMP-1, (b) N4CMP-2, (c) N4CMP-3, (d) N4CMP-4, and (e) N4CMP-5.

Full size image

The porosity of the polymer networks was investigated by nitrogen adsorption/desorption experiments at 77 K, in which the fully reversible isotherms exhibit rapid nitrogen uptake at low relative pressures (P/P 0 < 0.01), indicates that the five polymers belong to microporous materials (Fig. 4a and Figure S6). Meanwhile, all the polymer networks gave rise to type I nitrogen gas sorption isotherms according to the IUPAC classification21. N4CMP-1, N4CMP-2 and N4CMP-3 show little hysteresis upon desorption, suggesting that adsorption and desorption are almost equally facile. Nevertheless, N4CMP-4 and N4CMP-5 networks show an evident hysteresis loop, which is partly attributed to the swelling in a flexible polymer network, as well as mesopore contribution34,35. In fact, the described low-pressure hysteresis is common phenomenon in a variety of microporous polymer networks18,19,36,37,38,39,40,41. The presence and magnitude of this hysteresis may be an indication of the softness of the material. Besides, a sharp rise in the high pressure region (P/P 0 > 0.8) is also observed in the sorption isotherms of N4CMP-1, N4CMP-2 and N4CMP-3, respectively, suggesting that the materials possess some macropores, which is attributed to the nitrogen condensation in interparticular voids formed by the aggregation of polymer microspheres observed in the SEM images, which is similar to the previous reports42,43. The apparent BET surface areas for the networks (Table S1) were calculated over a relative pressure range P/P 0 = 0.015–0.1, which was found to give a positive value of C in the BET equation. Polymer N4CMP-3 shows the highest BET surface areas with a value of 1426 m2 g−1, while N4CMP-2 represents the lowest value of 592 m2 g−1 among the five polymer networks. As a general trend, when the similar synthetic methodology was used, polymers prepared from longer linkers showed lower surface areas than those from shorter linkers, the similar results were reported by the Cooper and Thomas group23,24,36. For example, N4CMP-1 had a BET surface area of 650 m2 g−1, which decreased to 592 m2 g−1 for N4CMP-2. This phenomenon is similar to that of reported previously, in which CMPs constructed with longer connecting struts have lower BET surface areas23,36. Beyond this, Jiang et al. reported the linkage geometry between the connecting nodes could affect pore properties, conjugation and surface area for CMP networks37. Their experiment results demonstrated that the meta-linkage is superior to the ortho- and para-linkages for the construction of porous skeletons. This phenomenon is also observed in this system, for example, polymer N4CMP-3 containing the meta-linkage 1,3,5-triethylbenzene possesses the highest surface area (1426 m2 g−1), which is higher than that of polymers N4CMP-4 (995 m2 g−1) and N4CMP-5 (1347 m2 g−1) with the ortho- and para-linkage. In addition, the polymer N4CMP-5 shows a higher surface area in the contrastion of N4CMP-4, possible reason is that the effect of steric hindrance from tetrakis(4-bromophenyl)methane improved the further growth of the molecular weight of the polymer during polymerization, which leads to the increase of surface area. The micropore volume was calculated from the adsoprtion branch of the nitrogen adsorption-desorption isotherm using the t-plot method (ESI, Figure S7). A summary of the BET specific surface area and porous properties is shown in Table S1.

Figure 4
figure 4

(a) Nitrogen sorption isotherms measured at 77 K for N4CMP-1–5, the isotherms of N4CMP-1–4 are shifted vertically by 800, 600, 400 and 200 cm3 g−1 for better visibility, respectively; (b) pore size distributions calculated using nonlocal density functional theory (NLDFT) method, for clarity, the curves of N4CMP-1–4 are shifted vertically by 4, 3, 2 and 1 cm3 g−1, respectively.

Full size image

The pore size distribution (PSD) of polymers N4CMPs was calculated from the related adsorption branch of the isotherms by the NLDFT method. As shown in Fig. 4b, the N4CMPs polymers based on different building blocks showed the dominant pore sizes range between 0.36 and 2.0 nm. Polymers N4CMP-1 and 2 possessing same core structures were synthesized by bis-substituted monomers with the increasing length (biphenylene, diethynylbenzene), and N4CMP-3–5 were prepared with different geometry linkage nodes. For instance, N4CMP-3 has smaller micropore diameter (0.36 nm) and higher surface area compared to other four polymers, this result could be explained by the fact that the meta-linkage is superior to the ortho- and para-linkages for the construction of porous skeletons, similar results were observed for other reported CMPs24,31,37. Meanwhile, from the N2 sorption isotherm, the ratio of micropore volume to total pore volume (V Micro/V Total) can be calculated which describes the degree of microposity. All the five polymers exhibited the V Micro/V Total values above 0.55, ranging from 0.57 to 0.79, which indicates that a high fraction and dominance of micropores in these networks. These results implyed that the BET surface area and porosity porosities in the CMP networks could be finely controlled by using either linker with different strut length or different geometries. The key structural properties of polymers derived from the corresponding isotherm are listed in Table S1, such as the BET specific surface area, micropore surface area, pore volume, micropore volume, dominant pore size, gas uptake, and selectivity.

Discussion

Gas uptake capacity and separation

It has been proved that compare to other most porous polymers, carbazole-based polymer networks can efficiently adsorb gas (hydrogen and carbon dioxide) under the same conditions32,38,39,40,41. Therefore, gas uptake capacities of the nitrogen-rich polymers were investigated. The CO2 uptakes of the polymer networks were studied at 1.05 bar and different temperature (273 K and 298 K) (Fig. 5a and Figure S8). We find that an increasing trend in the carbon dioxide loading capacity (from 2.05 mmol g−1 to 3.62 mmol g−1) with increasing micropore surface area and pore volume (Table S1). N4CMP-3 with the highest micropore surface area and micropore volume, shows the highest CO2 uptake of 3.62 mmol g−1 at 273 K and 1.05 bar among the resulting polymer networks, followed by N4CMP-5 (3.18 mmol g−1), N4CMP-4 (2.49 mmol g−1), N4CMP-1 (2.24 mmol g−1), and N4CMP-2 (2.05 mmol g−1) (Table S2 in the Supporting Information). From this result, we found that N4CMP-1 and N4CMP-2 have similar micropore surface area, however, N4CMP-1 shows the relatively higher CO2 uptake, possible reason is that N4CMP-1 has high micropore volume (Table S1). Interestingly, these CO2 uptake values are not only significantly higher than many microporous materials with a similar specific surface area but also comparable to the reported large surface area of porous aromatic frameworks under the same conditions, such as PAF-1 (2.1 mmol g−1, S BET = 5640 m2 g−1)44, COF-102 (1.56 mmol g−1, S BET = 3620 m2 g−1)45, and PP-CMP@mmm (2.52 mmol g−1, S BET = 1928 m2 g−1)37, but lower than TSP-2 (4.1 mmol g−1, S BET = 913 m2 g−1)40, ALP-1 (5.4 mmol g−1, S BET = 1235 m2 g−1)46 and PPF-1 (6.1 mmol g−1, S BET = 1740 m2 g−1)47, implying that the surface area of porous polymers is not a sole factor in determining the capacity of CO2 uptake. The superior adsorption properties of the N4CMP polymers can be ascribed to the enhanced dipole-quadrupole interactions and/or the strong interactions of the polarizable CO2 molecules through hydrogen bonding (from the N-H group), similar results were also observed for previously reported carbazole-based CMPs23. The recycling is an important parameter for their practical application. We tested the reusability of N4CMP-3, N4CMP-4 and N4CMP-5 at 1.05 bar and 273 K, and found the samples could be efficiently recycled and reused for four cycles without significant loss of CO2 uptake (ESI, Figure S9).

Figure 5
figure 5

CO2 adsorption isotherms collected (a) at 273 K and 1.05 bar and (b) at 318 K and 50 bar, respectively; (c) The isosteric heat of adsorption for N4CMPs.

Full size image

Furthermore, a high pressure CO2 sorption properties of the polymers were also investigated at 50 bar and 318 K. As seen in Fig. 5b, N4CMP-2 and N4CMP-3 produce a type IV isotherm according to IUPAC classifications48. N4CMP-1, N4CMP-4 and N4CMP-5 show a nearly linear increase with the increasing pressure no obviously turning point. All the polymers show similar growth trend in the capture of carbon dioxide gas (from 542 mg g−1 to 800 mg g−1) with increasing surface area and pore volume under high pressure condition. N4CMP-3 exhibits the highest CO2 capture capacity of 800 mg g−1. N4CMP-1, N4CMP-2, N4CMP-4, and N4CMP-5, exhibit the CO2 uptake of 605, 542, 556, and 695 mg g−1, respectively. These results indicate that electron-rich polymer backbone, porosity property and pressure play a positive role in the increase of gas capature capacity.

To further understand the pore surface properties and the adsorption process, the isosteric heat (Q st) of polymers N4CMP for adsorption CO2 is calculated based on adsorption isotherms of CO2 at 273 K and 298 K in terms of the Clausius–Clapeyron equation (Fig. 5c)49. The Q st values of CO2 drop with loading amount, meaning that the interaction between CO2 and the pore wall is stronger than that between CO2 molecules. The Q st values of CO2 adsorption for all the polymer networks range from 25.5 to 35.1 kJ mol−1 at the near zero coverage, which can be comparable to other carbazole-based polymers and higher than those of some reported POPs, such as polycarbazole CPOP-1 (24.5–30.2 kJ mol−1)38,40,41, HCP materials (20–24 kJ mol−1)50, polybenzimidazole BILP polymers (26.7–28.8 kJ mol−1)51. The higher Q st value may be attributed to the appropriate pore structures for CO2 adsorption, electron-rich polycarbazole network and high charge density at the nitrogen sites of polymers that can enhance the interaction between CO2 molecules and the pore surfaces of N4CMP.

Besides gas storage, the selectivity is very important for potential application in gas separation. Considering the good gas adsorption performance of the obtained polymers, the gas selective adsorption behaviors of N4CMP polymers are measured. The sorption experiments of CO2, CH4, and N2 were carried out at 273 K and 1.05 bar, respectively. The CO2 or CH4 uptake shows a almost linear increase with the increasing pressure, whereas that of nitrogen has no apparent increase trend (ESI, Figure S10). From the available single-gas adsorption isotherms, the selectivity in the adsorption of CO2, CH4 and N2 from CO2-N2, CO2-CH4, and CH4-N2 gas mixtures was estimated by Ideal Adsorption Solution Theory (IAST), which has been widely used to predict gas mixture adsorption behavior in the porous materials52,53. Under simulated flue gas streams (typically 15% CO2 and 85% N2), N4CMP-3 exhibits the highest selectivity for CO2/N2 among the five polymers (53.8 at 273 K and 1.05 bar) (Fig. 6a). This value is lower than that of porous benzimidazole linked polymers BILP 1–7 (59‒113)54 and PECONF 1‒4 (74‒109)55, while it is comparable to those of previously reported carbazole-based polymers (29.7‒35.4)31. In addition, the CO2/CH4 adsorption selectivity for N4CMP polymer networks is calculated to be 4.6–5.2 at 273 K and 1.05 bar, which is moderate value (Fig. 6b and Table S1). The CH4/N2 selectivity for N4CMP polymers is 5.7–10.5 (Fig. 6c and Table S1). Furthermore, we employ initial slopes ratio estimated from Henry’s law constants for single-component adsorption isotherms, which has been widely adopted by researchers56. the CO2 selectivities of N4CMP-1‒5 over N2 and CH4 were also calculated using initial slopes calculations at 273 K and 1.0 bar, the selectivities of CO2/N2 (46.2–69.1) (ESI, Figure S11, Table S1) and CO2/CH4 (6.4–12.1) (ESI, Figure S12, Table S1). The CH4/N2 adsorption selectivities range from 4.0 to 10.6 (ESI, Figure S13, Table S1). These excellent CO2 selective capture performances of N4CMP samples evaluated by Henry’s law are consistent with the results calculated from IAST. Due to their relatively high heats of adsorption and good selectivity of CO2/N2, CO2/CH4 and CH4/N2, the N4CMPs synthesized in this work could have potential for post-combustion CO2 capture.

Figure 6
figure 6

IAST selectivity for 0.15/0.85 CO2/N2 mixture, 0.50/0.50 CO2/CH4 mixture, and 0.50/0.50 CH4/N2 mixture.

Full size image

Conclusion

In conclusion, a series of carbazole-containing CMPs have been synthesized by palladium (0) catalyzed cross-coupling polycondensation. The study suggests that the monomer reactivity and the efficacy of chemistry have a pronounced effect on the surface areas and porosity attained in addition to the monomer geometry structure and linker strut length. All of the polymers are microporous with the BET surface areas ranging from 592 to 1426 m2 g−1 and possess quite narrow pore distributions at around 0.36 and 0.61 nm. The thermally stable polymer N4CMP-3 exhibits the highest CO2 adsorb uptake up to 3.62 mmol g−1 (1.05 bar/273 K) and 800 mg g−1 (50 bar/318 K) among the obtained polymers. All of the CMP networks show high isosteric heats of CO2 adsorption (25.5–35.1 kJ mol−1). We believe that this type of CMPs material would be a promising candidate for clean energy application and environmental-friendly field.

Methods

Synthesis of N4CMP-1

A mixture of 1,3,6,8-tetrabromocarbazole (100 mg, 0.21 mmol) and biphenyl-4,4′-diboronic acid (101.5 mg, 0.42 mmol) in dimethylformamide (DMF, 8 mL) was degassed by three freeze–pump–thaw cycles. To the mixture was added an aqueous solution of K2CO3 (1.0 M, 1 mL) and tetrakis(triphenylphosphine)palladium (0) (14.8 mg, 12.8 μmol), respectively. The resulting solution was further degassed for three cycles, and purged with N2, and stirred at 120 °C for 48 h. The mixture was cooled to room temperature, and then insoluble precipitate was filtered and washed with H2O, CH3OH, CHCl3, and THF to remove any unreacted monomers or catalyst residues. Further purification of the polymer was carried out by Soxhlet extraction with H2O, CHCl3, THF, and CH3OH for 24 h, respectively, N 4 CMP-1 (gray powder, 93 mg, 94.7% yield). Elemental Analysis (%) (N 4 CMP-1) C 92.86, H 5.37, N 1.77; Found C 91.11, H 5.10, N 2.06.

Synthesis of N4CMP-2, N4CMP-3, N4CMP-4, and N4CMP-5

1,3,6,8-Tetrabromocarbazole (100.0 mg, 0.21 mmol) and 1,4-diethynylbenzene (52.9 mg, 0.42 mmol) (N 4 CMP-2)/1,3,5-triethynylbenzene (42 mg, 0.28 mmol) (N 4 CMP-3)/1,1′,2,2′-tetrakis(4-ethynylphenyl)ethene (90 mg, 0.21 mmol) (N 4 CMP-4)/tetrakis(4-ethynylphenyl)methane (87.5 mg, 0.21 mmol) (N 4 CMP-5) were put into a 50 mL round-bottom flask, then the flask exchanged 3 cycles under vacuum/N2. Then added to 2 mL DMF and 2 mL triethylamine (Et3N), the flask was degassed by three freeze–pump–thaw cycles, purged with N2. When the solution had reached reaction temperature, a slurry of tetrakis(triphenylphosphine)palladium (0) (14.5 mg, 12.6 μmol) (N 4 CMP-2)/(N 4 CMP-3)/(N 4 CMP-4)/(N 4 CMP-5) in the 1 mL DMF and copper (I) iodide (3.8 mg, 0.02 mmol) (N 4 CMP-2)/(N 4 CMP-3)/(N 4 CMP-4)/(N 4 CMP-5) in the 1 mL Et3N were added, and the resulting solution was further degassed by freeze-pump-thaw for three cycles, and purged with N2, and stirred at 120 °C for 48 h, respectively. The mixture was cooled to room temperature, and then insoluble precipitate was filtered and washed with H2O, CH3OH, CHCl3, and THF to remove any unreacted monomers or catalyst residues. Further purification of the polymer was carried out by Soxhlet extraction with H2O, CHCl3, THF, and CH3OH for 24 h, respectively, to give N 4 CMP-2 (yellow solid, 81 mg, 94.5%), N 4 CMP-3 (brown solid, 71 mg, 95%), N 4 CMP-4 (black solid, 113 mg, 92%), N 4 CMP-5 (brownish black solid, 109 mg, 90.6%). Elemental Analysis (%) Calcd. (N 4 CMP-2) C 94.24, H 3.65, N 2.11; Found C 92.88, H 3.87, N 2.44; (N 4 CMP-3) C 94.97, H 3.19, N 1.84; Found C 92.78, H 3.48, N 2.06; (N 4 CMP-4) C 95.51, H 3.84, N 0.65; Found C 93.84, H 4.05, N 0.98; (N 4 CMP-5) C 94.81, H 4.42, N 0.77; Found C 92.84, H 4.75, N 1.04.