Porous Organic Polymers Derived from Ferrocene and Tetrahedral Silicon-Centered Monomers for Carbon Dioxide Sorption

Herein, we present two novel ferrocene-containing porous organic polymers, FPOP-1 and FPOP-2, by the Heck reactions of 1,1′-divinylferrocene with two tetrahedral silicon-centered units, i.e., tetrakis(4-bromophenyl)silane and tetrakis(4′-bromo-[1,1′-biphenyl]-4-yl)silane. The resulting materials possess high thermal stability and moderate porosity with the Brunauer–Emmer–Teller (BET) surface areas of 499 m2 g−1 (FPOP-1) and 354 m2 g−1 (FPOP-2) and total pore volumes of 0.43 cm3 g−1 (FPOP-1) and 0.49 cm3 g−1 (FPOP-2). The porosity is comparable to previously reported ferrocene-containing porous polymers. These materials possess comparable CO2 capacities of 1.16 mmol g−1 (5.10 wt%) at 273 K and 1.0 bar, and 0.54 mmol g−1 (2.38 wt%) at 298 K and 1.0 bar (FPOP-1). The found capacities are comparable to, or higher than many porous polymers having similar or higher surface areas. They have high isosteric heats of up to 32.9 kJ mol−1, proving that the affinity between the polymer network and CO2 is high, which can be explained by the presence of ferrocene units in the porous networks. These results indicate that these materials can be promisingly utilized as candidates for the storage or capture of CO2. More ferrocene-containing porous polymers can be designed and synthesized by combining ferrocene units with various aromatic monomers under this strategy and their applications could be explored.


Introduction
Porous organic polymers (POPs) are a specific class of covalently bonded functional materials containing permanent porosity and constructed from organic monomers through various synthetic methodologies. By virtue of their diversified functionality (e.g., porosity, fluorescence, ionic charges), POP materials have found extensively potential applications in gas separation, gas storage, capture of pollutants, optoelectronics, energy storage, catalysis, sensing, nanoreactors, and medical applications [1][2][3][4][5][6]. Among various POP materials, the metal atoms containing POPs have gained particular interest as the incorporation of metal sites into the porous networks could impart the resultant materials new functions (e.g., catalytic activity and fluorescence), and thus extend/enhance their application performance (e.g., enhanced CO 2 adsorption) [7][8][9][10][11][12][13][14][15][16]. For example, Son et al., reported a mesoionic carbene-Rh species containing microporous organic polymer exhibiting high catalytic activity in the stereoselective polymerization of arylacetylenes by the functionalization of azide-alkyne click-based microporous polymer [9]. The vanadium-docked covalent organic frameworks have served as efficient heterogeneous catalysts for Mannich-type reactions [17]. By doping alkali-metal into POP networks, the CO 2 and H 2 uptakes could angle-spinning (CP/MAS) NMR and 29 Si MAS NMR spectra were measured by a Bruker AVANCE-500 NMR Spectrometer, while the operating magnetic field strength is 9.4 T. The field-emission scanning electron microscopy (FE-SEM) images was observed by a FEI QUANTA FEG250 Spectrometer (FEI, Hillsboro, OR, USA). Before observation, the samples were prepared by dropping the suspensions in ethanol on a copper grid and following air drying. The thermogravimetric analysis (TGA) was determined by a MettlerToledo SDTA-854 TGA system (Greifensee, Switzerland). The samples were heated at a rate of 10 • C min −1 in the range of room temperature to 800 • C in an atmosphere of nitrogen.
Nitrogen sorption isotherms were measured at 77 K with a Micromeritics ASAP 2020 (Norcross, GA, USA) volumetric adsorption analyzer. Prior to the measurement, the samples of ca. 100 mg were degassed in vacuum at 150 • C for 12 h. The specific surface areas were calculated based on the nitrogen adsorption data by Brunauer-Emmett-Teller (BET) analysis in the relative pressure (P/P 0 ) range of 0.01~0.20, whereas the pore size distribution was evaluated on the basis of nonlocal density functional theory (NL-DFT). The total pore volumes were calculated at P/P 0 = 0.99, whereas the micropore surface area and micropore volume were calculated using the t-plot method. The carbon dioxide (CO 2 ) adsorption isotherms were measured on a Micrometrics ASAP 2020 volumetric adsorption analyzer at 298 K and 273 K. Before the measurements, the samples were degassed in vacuum at 150 • C for at 12 h.

Synthesis of FPOP-1
In a three-necked bottle, 1,1 -divinylferrocene (0.24 g, 1 mmol), tetrakis(4-bromophenyl) silane (0.33 g, 0.5 mmol), K 2 CO 3 (0.55 g, 4 mmol), Pd(PPh 3 ) 4 (0.12 g, 0.10 mmol), and 20 mL of freshly distilled DMF were added under an atmosphere of argon. The argon was bubbled into the mixture for at least 0.5 h to remove oxygen in the system. Then, the mixture was heated to 120 • C and kept for 72 h under argon. The resultant mixture was cooled to room temperature and filtered with a vacuum equipment. The resultant precipitate was washed with various solvents, including water, methanol, chloroform, tetrahydrofuran (THF), and acetone to remove the residuals. The obtained crude product was further purified by Soxhlet extraction with methanol and THF for 24 h, respectively. Finally, the product, which was denoted as FPOP-1, was dried in a vacuum oven at 80 • C for 24 h and appeared as a red-brown solid (213 mg, yield: 53.0%). Elemental analysis calc. (wt%) for SiC 52 H 36 Fe 2 : C 78.2, H 4.3; found C 72.5, H 4.0.

Synthesis and Characterization
The synthetic routes of two novel ferrocene-based porous organic polymers are shown in Scheme 1. FPOP-1 and FPOP-2 were prepared by the Heck coupling reactions of 1,1 -divinylferrocene with two tetrahedral silanes, i.e., tetrakis(4-bromophenyl)silane and tetrakis(4 -bromo-[1,1 -biphenyl]-4-yl)silane. The reactions were conducted using Pd(PPh 3 ) 4 as the catalyst and K 2 CO 3 as the acid absorbent in DMF at 120 • C for 72 h [40]. Upon the completion of the reaction, the crude products were filtrated and washed with various solvents to remove the residuals, including unreacted monomers, residual catalysts, and other impurities in the final products. Subsequently, the products were further purified by Soxhlet extraction and dried in a vacuum at 80 • C for 24 h. The final materials were afforded as red-brown powders. As expected, these polymers are insoluble in common solvents such as tetrahydrofuran, water, ethanol, and DMF. various solvents to remove the residuals, including unreacted monomers, residual catalysts, and other impurities in the final products. Subsequently, the products were further purified by Soxhlet extraction and dried in a vacuum at 80 °C for 24 h. The final materials were afforded as red-brown powders. As expected, these polymers are insoluble in common solvents such as tetrahydrofuran, water, ethanol, and DMF. To confirm the structures of these polymers, they were fully characterized by FT-IR, solid-state 13 C CP/MAS NMR, and 29 Si NMR spectroscopy. FPOP-1 and FPOP-2 display similar spectra due to their similar chemical structures. For FT-IR spectra as shown in Figure 1, the multiple peaks from 1600 cm −1 to 1400 cm −1 are obviously attributed to the C=C stretching vibrations from phenyl and cyclopentadiene rings in tetrahedral siliconcentered units and ferrocene units. The peaks ranging from 3000 cm −1 to 3100 cm −1 are apparently associated to the unsaturated C-H stretching vibrations from phenyl and cyclopentadiene rings. In addition, two kinds of unexpected peaks were observed at ca. 3450 cm −1 and 2900-3000 cm −1 . The former was attributed to -OH groups from water encapsulated in the samples, while the latter was aroused from saturated C-H groups, which are from the residual solvents (e.g., THF and methanol) in the network during the Soxhlet extraction process. Although the samples have been dried under vacuum, it is very difficult to completely remove all the residual solvents, including water, THF and methanol, because some of them may be embedded in the closed pores. This finding is consistent with our previous report [41]. To confirm the structures of these polymers, they were fully characterized by FT-IR, solid-state 13 C CP/MAS NMR, and 29 Si NMR spectroscopy. FPOP-1 and FPOP-2 display similar spectra due to their similar chemical structures. For FT-IR spectra as shown in Figure 1, the multiple peaks from 1600 cm −1 to 1400 cm −1 are obviously attributed to the C=C stretching vibrations from phenyl and cyclopentadiene rings in tetrahedral siliconcentered units and ferrocene units. The peaks ranging from 3000 cm −1 to 3100 cm −1 are apparently associated to the unsaturated C-H stretching vibrations from phenyl and cyclopentadiene rings. In addition, two kinds of unexpected peaks were observed at ca. 3450 cm −1 and 2900-3000 cm −1 . The former was attributed to -OH groups from water encapsulated in the samples, while the latter was aroused from saturated C-H groups, which are from the residual solvents (e.g., THF and methanol) in the network during the Soxhlet extraction process. Although the samples have been dried under vacuum, it is very difficult to completely remove all the residual solvents, including water, THF and methanol, because some of them may be embedded in the closed pores. This finding is consistent with our previous report [41].  Figure 2 shows the solid-state 13 C CP/MAS NMR spectra of these materials and the resonances were assigned. As expected, the peaks attributable to sp 2 carbon atoms from ethenylene, phenylene and cyclopentadiene groups in ferrocene and silicon-centered units were observed with overlapping from 120 ppm to 150 ppm. For 29 Si NMR spectra, both of FPOP-1 and FPOP-2 display one peak at −15.5 ppm and −15.1 ppm, respectively   Figure 2 shows the solid-state 13 C CP/MAS NMR spectra of these materials and the resonances were assigned. As expected, the peaks attributable to sp 2 carbon atoms from ethenylene, phenylene and cyclopentadiene groups in ferrocene and silicon-centered units were observed with overlapping from 120 ppm to 150 ppm. For 29 Si NMR spectra, both of FPOP-1 and FPOP-2 display one peak at −15.5 ppm and −15.1 ppm, respectively (Figure 3), which are unambiguously derived from the Si atoms from tetrahedral Si(p-C 6 H 4 ) 4 or Si(p-C 6 H 4 -C 6 H 4 ) 4 units. This finding indicates that the tetrahedral silicon units are chemically stable and no Si-C bonds have been cleaved during the reaction, in contrary to the possible Si-C cleavage under base conditions (e.g., triethylamine or K 2 CO 3 as the acid adsorbents) reported in previous literatures [38,42]. These results are in accordance with the 29 Si NMR results found in silicon-centered POPs [43,44].  Figure 2 shows the solid-state 13 C CP/MAS NMR spectra of these materials and the resonances were assigned. As expected, the peaks attributable to sp 2 carbon atoms from ethenylene, phenylene and cyclopentadiene groups in ferrocene and silicon-centered units were observed with overlapping from 120 ppm to 150 ppm. For 29 Si NMR spectra, both of FPOP-1 and FPOP-2 display one peak at −15.5 ppm and −15.1 ppm, respectively (Figure 3), which are unambiguously derived from the Si atoms from tetrahedral Si(p-C6H4)4 or Si(p-C6H4-C6H4)4 units. This finding indicates that the tetrahedral silicon units are chemically stable and no Si-C bonds have been cleaved during the reaction, in contrary to the possible Si-C cleavage under base conditions (e.g., triethylamine or K2CO3 as the acid adsorbents) reported in previous literatures [38,42]. These results are in accordance with the 29 Si NMR results found in silicon-centered POPs [43,44].    Figure 4 shows the filed-emission scanning electron microscopy (FE-SEM) images of FPOP-1 and FPOP-2. As expected, these polymers are amorphous and exhibit irregular spherical or bulk particles with the scale at the micrometer sizes.    Figure 4 shows the filed-emission scanning electron microscopy (FE-SEM) images of FPOP-1 and FPOP-2. As expected, these polymers are amorphous and exhibit irregular spherical or bulk particles with the scale at the micrometer sizes.  Figure 5 shows the thermogravimetric analysis (TGA) spectra of these two polymers and the analysis was conducted in an atmosphere of nitrogen at a heat rate of 10 °C min −1 ranging from room temperature to 800 °C. Similar to other POP materials [43,44], these polymers also have good thermal stability, while the decomposition temperatures at 5% mass loss (Td,5%) were found at 405 °C and 368 °C for FPOP-1 and FPOP-2, respectively. In addition, after the decomposition, the residual ratio of these polymers was higher than 65% at 800 °C. This finding is obviously attributed to their highly cross-linking networks.   Figure 5 shows the thermogravimetric analysis (TGA) spectra of these two polymers and the analysis was conducted in an atmosphere of nitrogen at a heat rate of 10 • C min −1 ranging from room temperature to 800 • C. Similar to other POP materials [43,44], these polymers also have good thermal stability, while the decomposition temperatures at 5% mass loss (T d,5% ) were found at 405 • C and 368 • C for FPOP-1 and FPOP-2, respectively. In addition, after the decomposition, the residual ratio of these polymers was higher than 65% at 800 • C. This finding is obviously attributed to their highly cross-linking networks.  Figure 4 shows the filed-emission scanning electron microscopy (FE-SEM) images of FPOP-1 and FPOP-2. As expected, these polymers are amorphous and exhibit irregular spherical or bulk particles with the scale at the micrometer sizes.  Figure 5 shows the thermogravimetric analysis (TGA) spectra of these two polymers and the analysis was conducted in an atmosphere of nitrogen at a heat rate of 10 °C min −1 ranging from room temperature to 800 °C. Similar to other POP materials [43,44], these polymers also have good thermal stability, while the decomposition temperatures at 5% mass loss (Td,5%) were found at 405 °C and 368 °C for FPOP-1 and FPOP-2, respectively. In addition, after the decomposition, the residual ratio of these polymers was higher than 65% at 800 °C. This finding is obviously attributed to their highly cross-linking networks.

Porous Properties
The porous properties of these polymers were evaluated by nitrogen adsorptiondesorption isotherms at 77 K ( Figure 6a) and the data are summarized in Table 1. The isotherms of FPOP-1 and FPOP-2 are similar. Both of them display both type I isotherm and type IV isotherm according to the IUPAC classification [45]. They give rise to a sharp capacity at relatively low pressure and a gradually increasing capacity at relatively higher pressure with hysteresis. This feature indicates that both micropores and mesopores exist within the network. On the basis of the N 2 isotherm, FPOP-1 has a moderate porosity with the Brunauer-Emmer-Teller (BET) surface area (S BET ) of 499 m 2 g −1 and the total pore volume (V total ) of 0.43 m 3 g −1 . The micropore surface area (S micro ) and the micropore volume (V micro ) were 308 m 2 g −1 and 0.23 cm 3 g −1 , respectively, calculated using the t-plot method. The V micro /V total ratio, which represents the contribution of microporosity in the framework, is 0.53, thereby indicating that mesopores and micropores provide nearly same contribution to the porous network. Compared with FPOP-1, FPOP-2 shows lower porosity with S BET of 354 m 2 g −1 and V total of 0.49 m 3 g −1 , and lower microporosity with S micro of 175 m 2 g −1 and V micro of 0.13 m 3 g −1 . As well, FPOP-2 possesses a much lower V micro /V total ratio of 0.27, thereby indicating that the mesopores mainly contribute to the network. The pore size distributions (PSDs) were estimated by nonlocal density functional theory (NL-DFT). As expected, the PSD results are in accordance with the shape of nitrogen isotherms. FPOP-1 displays two main peaks in the region of micropores centered at ca. 1.4 nm and mesopores centered at ca. 2.1 nm (Figure 6b). The PSD curve of FPOP-2 is similar to that of FPOP-1 except that the mesopore region is broader and centered at 2.8 nm. These results further reveal the co-existence of micropores and mesopores in the porous networks. ume (Vtotal) of 0.43 m 3 g −1 . The micropore surface area (Smicro) and the micropore volume (Vmicro) were 308 m 2 g −1 and 0.23 cm 3 g −1 , respectively, calculated using the t-plot method. The Vmicro/Vtotal ratio, which represents the contribution of microporosity in the framework, is 0.53, thereby indicating that mesopores and micropores provide nearly same contribution to the porous network. Compared with FPOP-1, FPOP-2 shows lower porosity with SBET of 354 m 2 g −1 and Vtotal of 0.49 m 3 g −1 , and lower microporosity with Smicro of 175 m 2 g −1 and Vmicro of 0.13 m 3 g −1 . As well, FPOP-2 possesses a much lower Vmicro/Vtotal ratio of 0.27, thereby indicating that the mesopores mainly contribute to the network. The pore size distributions (PSDs) were estimated by nonlocal density functional theory (NL-DFT). As expected, the PSD results are in accordance with the shape of nitrogen isotherms. FPOP-1 displays two main peaks in the region of micropores centered at ca. 1.4 nm and mesopores centered at ca. 2.1 nm (Figure 6b). The PSD curve of FPOP-2 is similar to that of FPOP-1 except that the mesopore region is broader and centered at 2.8 nm. These results further reveal the co-existence of micropores and mesopores in the porous networks.      Compared to previously reported ferrocene-containing porous polymers, the present polymers exhibit comparable porosity [28,29,46,47]. For example, Yamamoto reactions of 1,1 -dibromoferrocene and 5,10,15,20-tetrakis(4-bromophenyl)porphyrin or tetra(p-bromoph enyl)methane resulted in two ferrocene-based conjugated microporous polymers (CMPs) with the S BET of 638 m 2 g −1 and 422 m 2 g −1 [31]. We prepared a ferrocene-based POP with the S BET of 542 m 2 g −1 through the Sonogashira reaction of 1,1 -dibromoferrocene and tetrakis(4-ethynylphenyl)silane [48]. In addition, FPOP-1 exhibits higher porosity than FPOP-2. This finding may be due to the different structure length of these two siliconcentered monomers. It was known that longer monomer structure length could result in lower surface areas due to greater interpenetration and more efficient space filling in the network, which was induced by increased degree of conformational freedom [49]. In the present study, tetrakis(4 -bromo-[1,1 -biphenyl]-4-yl)silane has a longer structure length than tetrakis(4-bromophenyl)silane, thereby resulting in lower porosity.

Carbon Dioxide Sorption
One important application of the porous materials is gas sorption. Herein, carbon dioxide (CO 2 ) was selected as a representative gas because CO 2 emission has played an important role in global warming and become a major concern for environment [1,50]. Moreover, the capture of CO 2 is important and highly beneficial to reach net-zero emissions in the second half of the century [51]. Various porous materials, including zeolites, porous carbon, and metal-organic frameworks, have been applied as adsorbents for the capture or storage of CO 2 in addition to other efficient techniques, such as membrane separation, chemical adsorption, and gas hydrates [52][53][54]. Among various porous materials, POPs are particularly promising thanks to their high porosity and facile tunability of their pore surface functionality (e.g., ferrocene units in the networks), which can play an important role in improving the interactions with CO 2 . Therefore, the CO 2 adsorption ability of FPOP-1 and FPOP-2 was evaluated. The isotherms are shown in Figure 7 and the data are summarized in Table 1.

Conclusions
Two ferrocene-containing porous organic polymers (FPOPs) were synthesized based on 1,1'-divinylferrocene and tetrahedral silicon-based monomers, i.e., tetrakis(4-bromophenyl)silane and tetrakis(4'-bromo-[1,1'-biphenyl]-4-yl)silane via the Heck reaction. The resulting materials exhibit high thermal stability and comparable porosity with the highest BET surface area of 499 m 2 g −1 and total pore volume of 0.49 cm 3 g −1 . The porosity is comparable to previous ferrocene-based POPs. For applications, these polymers display CO2 uptakes of 1.16 mmol g −1 (5.10 wt%) at 273 K and 1.0 bar, and 0.54 mmol g −1 (2.38 wt%) at 298 K and 1.0 bar (FPOP-1), which are comparable to many POP materials with similar or higher surface areas. Moreover, they both possess high isosteric heats of up to 32.9 kJ mol −1 , indicating their high affinity towards CO2 due to the presence of ferrocene units in the porous networks. These results demonstrate that these materials can be promisingly used as adsorbents for the storage and capture of CO2. Further investigation will focus on the construction of ferrocene-based POPs with high surface areas by altering different aromatic monomers to react with 1,1'-divinylferrocene.  FPOP-1 possesses moderate CO 2 capacities of 1.16 mmol g −1 (5.1 wt%) at 273 K and 1.0 bar, and 0.54 mmol g −1 (2.38 wt%) at 298 K and 1.0 bar. Compared to FPOP-1, FPOP-2 with a lower surface area expectedly displays lower CO 2 capacity, which was found to be 1.06 mmol g −1 (4.66 wt%) at 273 K and 1.0 bar, and 0.52 mmol g −1 (2.29 wt%) at 298 K and 1.0 bar (Figure 7a). Although the values are not very high, they are comparable to some polymers having similar or much larger specific surface areas, such as triazinecontaining CMP (TCMP-5, S BET : 494 m 2 g −1 , 1.22 mmol g −1 at 273 K/1 bar) [55], CMP-5 (S BET : 494 m 2 g −1 , 1.22 mmol g −1 at 273 K/1 bar) [55], CMP-1-(NH 2 ) (S BET : 710 m 2 g −1 , 1.56 mmol g −1 at 273 K/1 bar), and COF-102 (S BET : 3620 m 2 g −1 , 1.56 mmol g −1 at 273 K/1 bar) [56]. The interactions between the polymer framework and CO 2 molecules were further evaluated by calculating the isosteric heats (Q st ) while employing the Clausius-Clapeyron equation from CO 2 isotherms at 273 K and 298 K. The Q st values of FPOP-1 and FPOP-2 are 32.9 kJ mol −1 and 32.5 kJ mol −1 at low CO 2 coverage, and gradually decrease to ca. 30 kJ mol −1 and 25 kJ mol −1 with the adsorption increasing at higher coverage, respectively (Figure 7b). These findings reveal that the polymer frameworks possess strong binding ability with CO 2 because of the presence of the ferrocene unit as a withdrawing electron group, which can strongly interact with electron-rich CO 2 molecules [24]. In addition, the values of Q st are comparable or higher than previous ferrocene-based POPs [44,47,48]. These results reveal that these polymers can be applied as efficient adsorbents for CO 2 capture and storage.

Conclusions
Two ferrocene-containing porous organic polymers (FPOPs) were synthesized based on 1,1 -divinylferrocene and tetrahedral silicon-based monomers, i.e., tetrakis(4-bromophen yl)silane and tetrakis(4 -bromo-[1,1 -biphenyl]-4-yl)silane via the Heck reaction. The resulting materials exhibit high thermal stability and comparable porosity with the highest BET surface area of 499 m 2 g −1 and total pore volume of 0.49 cm 3 g −1 . The porosity is com-Polymers 2022, 14, 370 9 of 11 parable to previous ferrocene-based POPs. For applications, these polymers display CO 2 uptakes of 1.16 mmol g −1 (5.10 wt%) at 273 K and 1.0 bar, and 0.54 mmol g −1 (2.38 wt%) at 298 K and 1.0 bar (FPOP-1), which are comparable to many POP materials with similar or higher surface areas. Moreover, they both possess high isosteric heats of up to 32.9 kJ mol −1 , indicating their high affinity towards CO 2 due to the presence of ferrocene units in the porous networks. These results demonstrate that these materials can be promisingly used as adsorbents for the storage and capture of CO 2 . Further investigation will focus on the construction of ferrocene-based POPs with high surface areas by altering different aromatic monomers to react with 1,1 -divinylferrocene.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.