Fully Conjugated Benzyne-Derived Three-Dimensional Porous Organic Polymers
Graphical Abstract
Synthesis of a 3D porous organic polymer starting from a simple 0D bistriflate precursor via in situ CsF catalyzed formation of benzyne. This is the first demonstration of 3D polymer being prepared from a single and simple precursor. The resulting benzyne-derived porous organic polymer showed 544 m2 g−1 surface area and IAST CO2/N2 selectivity of 65.4
Abstract
Porous organic polymers (POPs) have gained tremendous attention owing to their chemical tunability, stability and high surface areas. Whereas there are several examples of fully conjugated two-dimensional (2D) POPs, three-dimensional (3D) ones are rather challenging to realize in the absence of structural templates. Herein, we report the base-catalyzed direct synthesis of a fully conjugated 3D POPs, named benzyne-derived polymers (BDPs), containing biphenylene and tetraphenylene moieties starting from a simple bisbenzyne precursor, which undergoes [2+2] and [2+2+2+2] cycloaddition reactions to form BDPs primarily composed of biphenylene and tetraphenylene moieties. The resulting polymers exhibited ultramicroporous structures with surface areas up to 544 m2 g−1 and very high CO2/N2 selectivities.
Introduction
Porous organic polymers (POPs) have attracted considerable attention owing to their permanent porosity, structural tunability, high chemical and physical stability.1 POPs with extended π-networks are of particular interest due to their potential in organic electronic devices such as chemical sensing,2 light-emitting diodes,3 organic photovoltaics,4 or as electrode materials in Li-ion batteries.5 A variety of synthetic strategies towards functional POPs have been reported including coupling reactions,6 condensation reactions,7 self-condensation8 and cycloadditions.9 However, obtaining conjugated 3D POPs often requires molecular level design by engaging 3D precursors such as tetraphenylene (TP). Driven to solve the shortcomings of 3D POPs in terms of conjugation, we recently demonstrated the design and synthesis of a fully sp2 hybridized 3D POP based on the acid-promoted cyclodeoxygenation of a 3D epoxy-functionalized POPs synthesized through Diels–Alder cycloaddition polymerization.10 The limitation of this approach, however, lies in the multi-step synthesis of the precursors. This is, indeed, one of the common challenges in the preparation of POPs containing cyclooctatetraene core in general.7, 10a, 11 Therefore, the synthesis of 3D fully sp2-hybridized POPs prepared from a single precursor would present a significant advance. One such precursor is a benzyne, which is a highly reactive intermediate and can be formed through the removal two substituents from an aromatic ring. Benzynes can be further derivatized through nucleophilic aromatic substitution, Diels–Alder cycloaddition or [2+2], [2+2+2] and [2+2+2+2] cycloaddition reactions enabling the formation of biphenylene (BP), triphenylene and TP moieties, respectively.12 Due to its instability and high reactivity, benzyne is commonly generated in situ from o-aryltrimethylsilyl triflates.12c, 13 Owing to its high reactivity, controlled self-polymerization of benzyne is very challenging14 and was only achieved using metal catalysts leading to the formation of short, nonporous oligomers.12c, 15 Herein, we report, for the first time, the preparation of a 3D fully conjugated benzyne-derived porous organic polymer (BDP) incorporating primarily BP and TP building blocks, prepared via CsF activated self-polymerization of a single precursor, that is 2,5-bis(trimethylsilyl)-1,4-phenylene bis(trifluoromethanesulfonate).13 The presence of CsF led to the in-situ formation of benzyne, which mainly undergoes [2+2] and [2+2+2+2] cycloaddition reactions to form BDPs containing BP and TP moieties (Scheme 1). The resulting BDPs showed surface areas over 500 m2 g−1 and very high CO2/N2 IAST selectivities up to 60.
Proposed structure of BDPs and their formation from a single precursor via in-situ generation of benzyne.
Results and Discussion
The synthesis of BDPs was carried out under solvothermal conditions using different solvents at different temperatures to find the optimal reaction conditions (Table S1). In order to ensure the scalability and repeatability of our approach, the polymerizations were carried out in a microwave (MW) reactor. Additionally, some of the BDPs have also been prepared in a sealed ampoule for comparison (Sample name: BDP-Solvent-(A)-Temperature, A is used to show samples prepared in ampoule).
In an effort to optimize the reaction conditions, we first tested common polar solvents such as dimethylformamide (DMF) and N-methyl-2-pyrrolidone (NMP) in the temperature range of 160–200 °C,10a however, the reaction yields were found to be rather low. These polymers showed poor thermal stabilities and turned out to be amorphous and nonporous (Table S1–S2, Figure S1–S3). Subsequently, solvents with higher boiling points such as nitrobenzene, phenyl ether and 1,2,4-trichlorobenzene (TChB) were tested. Among these solvents, TChB was chosen owing to the consistently higher Brunauer, Emmett and Teller (BET) surface areas of the resulting BDPs.16 In an effort to identify the optimum reaction temperature, the polymerization reaction was performed at 150, 175, 200 and 250 °C in TChB. We observed rather low yields (<20 %) and the lowest BET surface area of 62.0 m2 g−1 for the reaction carried out at 150 °C (Table S2 and Figure S1a–S2). The BET surface areas of BDPs gradually increased with the reaction temperature up to 200 °C. Further increase in the reaction temperature to 250 °C led to BDP with a lower surface area. Importantly, thermal stability of the BDPs increased significantly compared to the ones obtained using DMF and NMP as solvents (Figure S3a). Except for the reaction carried out at 150 °C, we consistently observed the formation of crystalline salt species such as Cs2SiF6. In an effort to remove salts, Soxhlet extraction was performed using methanol. Although salts were no longer visible in the PXRD patterns (Figure S4), the TGA analysis revealed residual mass under the air atmosphere (Figure S5), which suggests that the presence of residual salts and inorganics inside the pores. In order to determine the salt content precisely, we performed inductively coupled plasma-optical emission spectroscopy (ICP-OES) analysis. The ICP-OES analysis revealed that Cs content is very low in the range of 13–17 ppm (Figure S6 and Table S3). PXRD analysis also revealed that BDPs were amorphous. The BET surface area of the activated polymer, BDP-TChB-200 was found to be 450 m2 g−1 (Figure S7 and Table S4).
The successful formation of the polymers was verified by Fourier-transform infrared spectroscopy (FTIR) analysis. FTIR spectra of BDPs revealed the significantly decreased intensity of −Si−CH3 and −C−F bands at ∼839 cm−1 and ∼1036 cm−1, respectively (Figure 1a and S8a). Our initial assumption regarding the formation of the BDPs involved a two-step process. First, the formation of an entirely one-dimensional (linear) polymer composed of BP units through [2+2] cycloaddition polymerization reaction (Scheme S1),17 which was followed by a transformation of the 1-D polymer into a 3D framework via interconversion of BP into TP upon thermal treatment at 400 °C.18 For this reason, BDPs were heated at 400 °C for 0.5 h under vacuum in an ampoule (Sample name: BDP-Solvent-(A)-Temperature-0.5). However, cross-polarization magic angle spinning nuclear magnetic resonance (CP-MAS NMR) (Figures 1b–c), FTIR (Figure S1b) and XRD (Figure S1b) analyses of BDPs after heating did not show clear signs of transformation. Most importantly, 13C CP-MAS NMR spectra of BDP-TChB-200 before and after the thermal treatment did not reveal fundamental changes in the aromatic region (Figure 1b–c) but did show a loss of terminal groups upon thermal treatment. XRD patterns also unveiled the formation of new salts after thermal treatment, which was attributed to the loss of terminal groups (Figures S1b and S8). Similarly, FTIR spectra also revealed the further loss of the end groups after heating verified by the disappearance of C−F and −Si−CH3 stretching bands at ∼1036 and ∼839 cm−1, respectively (Figure 1a). Moreover, differential scanning calorimetry (DSC) analysis (Figure S3b) also did not indicate a clear phase transformation. The analysis of CP-MAS 13C NMR spectra of BDPs before and after heat treatment revealed, contrary to our initial assumption, the co-existence of BP and TP units in the BDP structure (Figure 1b–c).
Spectroscopic analysis of the bistriflate precursor and BDPs under various conditions. a) FTIR spectra and b) 13C Multi-CP-MAS NMR spectra taken at MAS rate of 24 kHz and c) constrained deconvolution of the BDP TChB-200 samples before and after the heating for various durations
Based on the 13C chemical shifts of BP and TP monomers (Figure S9), the most deshielded 13C signal observed in BDP-TChB-200 spectra at 154 ppm and the shoulder at 121 ppm are attributed to BP component (Figure 1b–c). The feature at 140 ppm is attributed to TP component. The BDP-TChB-200 spectra shows a 13C resonance at 0 ppm which is attributed to unreacted TMS groups, while the features in the range of 50–80 ppm are typical of −CF3 groups on triflate19 and are attributed to terminal −OTf groups. The feature at 130 ppm shifts to a lower chemical shift as the BPD samples are annealed for a longer period of time, which could be indicative of a greater ratio of TP, however the intensity of 13C signal 140 ppm decreased slightly after annealing for 0.5 h. This feature is attributed to the C atom attached to the terminal −OTf moiety that overlaps with solid-state spectra of the TP monomer.
After the initial annealing for 0.5 h, the features of the spectrum remain largely unchanged except for a decrease in the −OTf and −TMS 13C signals at longer annealing times as well as the appearance of a small feature at 112 ppm in the alkene region (Figure 1b–c). An attempt at deconvoluting the spectra was made and the results are shown in Figure 1c, however, TP:BP ratios were not sufficiently consistent to be taken quantitatively. However, a qualitative assessment of changes in the spectral features between the BDP samples can be made. The 13C chemical shifts of BP and TP were also obtained in solution where 151.4 ppm corresponds to the BP monomer carbons labeled (see Figure S9): 2, 3, 7, 8, 128.2 ppm: 5, 6, 10, 11 and 117.3 ppm: 1, 4, 9, 12. The corresponding 13C chemical shifts of biphenylene monomer in the solid phase (color coded in Figure 1b) are: 151.5 ppm (orange), 129.2 ppm (blue), 118.7 ppm (red). The 13C chemical shifts of tetraphenylene in solution (Figure S10): 141.5 ppm (1–8), 129.0 ppm (9, 12, 13, 16, 17, 20, 21, 24), 127.2 ppm (10, 11, 15, 18, 19, 22, 23). The 13C chemical shifts of tetraphenylene in the solid phase (Figure 1b): 4 distinct peaks at 140–142 ppm (purple), broad peaks around 127–130 ppm (pink and gray).20
Quantitative 1H solid-state NMR spectra of BDPs are shown in Figure S11. The aromatic 1H feature at 6.9 ppm broadened as the annealing time of the BDPs increased, suggesting a larger distribution of environments experienced by aromatic 1H nuclei in the BDPs and the increased number of defects upon removal of terminal groups. A decrease in ratio TMS (0 ppm) to the peak at 6.9 ppm was observed as the annealing time increased, likely due to a removal of terminal groups. A broad feature at 3.5 ppm in BDP-TChB-200 could be indicative of moisture in the sample present before annealing. A feature at 1.24 ppm and 4.8 ppm could be attributed to the end groups. In order to understand the effect of the thermal treatment on the polymer structure, Raman spectra of BDP-TChB-200 before and after heat treatment were recorded (Figure S12). Raman spectra showed that the ID/IG ratio increased from 0.69 to 0.78 at 0.5 h of thermal treatment at 400 °C and to 0.80 at 1.0 h.21 The ID/IG ratio continued to increase to 1.06 after 2.0 h of thermal treatment at 400 °C, which indicates increased number of defects as also suggested by the 1H solid-state NMR analysis of BDPs with the larger distribution of environments (Figure S11). Notably, the BET surface area of BDP-TChB-200 increased to 544 m2 g−1 from 450 m2 g−1 owing to the thermal removal of terminal groups after the heating (Table 1 and Figure 2a). All the BDPs showed Type I isotherms, an indication for a highly microporous structure, along with a H4 hysteresis in the desorption branch pointing to the co-existence mesopores.
Sample |
BET[a] |
Smicro[b] |
Sext[c] |
Vtotal[d] |
Vmicro[e] |
Vext[f] |
|---|---|---|---|---|---|---|
|
[m2 g−1] |
[m2 g−1] |
[m2 g−1] |
[cm3 g−1] |
[cm3 g−1] |
[cm3 g−1] |
BDP-TChB-200 |
449.4 |
253.0 |
196.4 |
0.31 |
0.11 |
0.20 |
BDP-TChB-200-0.5 |
544.3 |
378.7 |
165.5 |
0.32 |
0.15 |
0.17 |
BDP-TChB-200-1.0 |
418.5 |
264.7 |
153.8 |
0.27 |
0.11 |
0.16 |
BDP-TChB-200-2.0 |
363.2 |
248.1 |
115.1 |
0.21 |
0.10 |
0.11 |
- [a] BET surface area calculated over the pressure range (P/P0) of 0.01–0.15. [b] Micropore surface area calculated using the t-plot method. [c] Sext=Stotal−Smicro. [d] Total pore volume obtained at P/P0=0.99. [e] Micropore volume calculated using the t-plot method. [f] Vext=Vtotal−Vmicro.
a) N2 adsorption-desorption isotherms at 77 K and b) NLDFT pore size distribution curves of the BDP-TChB-200 samples before and after the heating calculated using carbon slit model. Filled and empty branches represent adsorption and desorption respectively
Notably, the size of the hysteresis loop decreased gradually with increasing thermal treatment time suggesting a decrease in the mesopore content. The micropore ratio increased from 56.3 to 69.6 % and the micropore volume improved by 36.4 % (Figure 3a and Table 1), which was attributed to the formation and/or opening of new pores during the thermal treatment.7 Pore size distribution (PSD) analysis of BDP-TChB-200 was performed by using the nonlocal density-functional theory (NLDFT) using carbon-slit model, which showed the presence of two types of pores: ultramicropores at 5.36 Å and micropores at 14 Å (Figure 2b). However, after thermal treatment, the ratio of ultra-micropores at 5.36 Å increased almost two-fold while the micropores at 14 Å decreased (Figure 2b) pointing to the formation of new pores, which could also be verified by the decrease in the mesopore content after heating (Table 1 and Figure 3a). Mesopore contents of BDP-TChB-200 were evaluated using Barrett-Joyner-Halenda (BJH) analysis.22 BJH analysis revealed the presence of mesopores between 3–10 nm and the mesopore volume decreased substantially after the thermal treatment (Figure S13). In order to further elaborate on the effect of the heating, the heating duration was increased to 1.0 and 2.0 h (sample name: BDP-TChB-200-t, where t is the duration of the heating). BET surface area started to decrease upon heating for more than 0.5 h (Table 1 and Figure 2a,) and both the amount of micropores and mesopores also decreased (Figure 3a). This result was attributed to the irreversible side reactions and to the blockage of the pores after the prolonged heating.23 These results also verify that the first heating step primarily involves the loss of terminal groups and result in further polymerization. The thermal stabilities of BDP-TChB-200 samples also substantially increased after thermal treatment confirming further polymerization (Figure S5).
(a) Change of BET surface area, micropore and mesopore areas with the heating time, (b) Change of different gas uptake capacities with the heating time, (c) Change of IAST CO2/gas selectivities with the heating time. The IAST CO2/N2 (d), CO2/CH4 (natural gas) (e) and CO2/CH4 (landfill gas) (f) selectivities of the BDP-TChB-200 samples before and after the heating for various durations.
Morphologies of BDP-TChB-200 before and after the heat treatment were evaluated by scanning electron microscopy (SEM) analysis. The SEM images of BDP-TChB-200 showed the presence of sheet-like structures (Figure S16a-d).11 However, after the heat treatment, the morphology transformed into a more bulk-like structure.7, 24 Densification of the polymer network was observed in the SEM micrographs of BDP-TChB-200-2.0 (Figure S16i–m). These results suggest that while at the early stages of thermal treatment, the loss of terminal groups and further polymerization enable the formation of abundant ultramicropores, prolonged heating leads to the blockage of the pores through the irreversible side reactions on the polymer structure (Figures S17–18 and Tables S6–7).
The presence of abundant ultramicropores in the structure of BDPs prompted us to test the affinity of environmentally relevant small gases. Accordingly, CO2, CH4 and N2 uptake capacities of BDP-TChB-200 were tested at 273 K before and after heat treatment (Figure S19). BDPs showed the highest affinity towards CO2, followed by CH4, and N2 (Table S8), which is expected owing to the abundance of ultramicropores.25 Whereas BDP-TChB-200 showed a CO2 uptake capacity of 1.43 (0.40 at 0.15 bar) mmol g−1, it increased to 1.66 (0.58 at 0.15 bar) mmol g−1 at 273 K, 1.1 bar after 0.5 h of heating (Figure 3b). In agreement with the trend in the BET surface areas, the CO2 uptake capacities started to decrease in samples heated for longer than 1 h and dropped to 1.29 (0.48 at 0.15 bar) mmol g−1 (Figure 3b). Whereas the residual salts can impact the CO2 affinity and uptake capacity,26 this effect is negligible owing to the very low salt content (<1.0 wt. %) as verified by the ICP-OES measurements.
The CH4 and N2 uptake also followed a similar trend. Since micropores are responsible for the high gas affinity, the gas uptake capacities nicely followed the micropore vs heating time plot in Figure 3a.27 Considering the high affinity of BDPs towards CO2, we investigated the CO2 separation performance under flue gas, CO2/N2 (15 : 85), natural gas (CO2/CH4, 5 : 95) and land fill gas (CO2/CH4, 50 : 50) conditions. Accordingly, gas adsorption isotherms were fitted with a dual-site Langmuir model (Figure S20) and CO2/gas selectivities were calculated using ideal adsorbed solution theory (IAST).28 The IAST CO2/N2 (post-combustion mixture) selectivity of BDP-TChB-200 turned out to be 11.2, however, after 0.5 h of heat treatment, it drastically increased up to 65.4 owing to the increased ultramicropore content (Figure 3d). The CO2/N2 selectivity followed the micropore vs heating time trend in Figure 3a and started to decrease upon heating for more than 0.5 h (Figure 3c), which was attributed to a decrease in the micropore content. BDP-TChB-200 also showed promising IAST CO2/CH4 natural gas and landfill gas selectivities (Figure 3e–f).6, 29 The CO2/CH4 selectivities also followed a similar trend with the micropore vs heating time in Figure 3a (Figure 3c).
Conclusion
We have shown the facile synthesis of a 3D fully conjugated porous organic polymer starting from a simple, 0D organic precursor through cycloaddition polymerization reaction of benzynes. The polymers showed high surface areas and exceptional gas selectivities. While it was unexpected, the formation and encapsulation of, otherwise moisture sensitive, Cs2SiF6 salt in an aromatic framework could potentially extend the applications of these polymers to light-emitting diodes.30 This work introduces a simple and efficient strategy for the preparation of fully conjugated 3D polymers towards gas and energy storage applications and is expected to further stimulate the research efforts for the preparation of fully conjugated all carbon 3D frameworks.
Supporting Information
The authors have cited additional references within the Supporting Information.25, 31 The Supporting Information is available free of charge on the. Materials and methods, synthetic procedures, supporting data including SEM, XRD, NMR, ICP-OES and Raman spectra. The data that support the findings of this study are openly available in Zenodo under doi. 10.5281/zenodo.7893679
Acknowledgments
A.C. acknowledges support from Swiss National Science Foundation (SNF) for funding of this research (200021-188572). Open Access funding provided by Université de Fribourg.
Conflict of interest
The authors declare no conflict of interest.
Open Research
Data Availability Statement
The data that support the findings of this study are openly available in ZENODO at https://doi.org/10.5281/zenodo.7893679, reference number 7893679.
