A Pressure Swing Approach to Selective CO2 Sequestration Using Functionalized Hypercrosslinked Polymers

Functionalized hypercrosslinked polymers (HCPs) with surface areas between 213 and 1124 m2/g based on a range of monomers containing different chemical moieties were evaluated for CO2 capture using a pressure swing adsorption (PSA) methodology under humid conditions and elevated temperatures. The networks demonstrated rapid CO2 uptake reaching maximum uptakes in under 60 s. The most promising networks demonstrating the best selectivity and highest uptakes were applied to a pressure swing setup using simulated flue gas streams. The carbazole, triphenylmethanol and triphenylamine networks were found to be capable of converting a dilute CO2 stream (>20%) into a concentrated stream (>85%) after only two pressure swing cycles from 20 bar (adsorption) to 1 bar (desorption). This work demonstrates the ease with which readily synthesized functional porous materials can be successfully applied to a pressure swing methodology and used to separate CO2 from N2 from industrially applicable simulated gas streams under more realistic conditions.


Introduction
The 2015 Paris Agreement aims to limit the average global temperature increase to 2 • C. One of the key causes of climate change is anthropogenic carbon dioxide, and recently, the UK government committed to a net zero emissions target by 2050. In the long term, the most effective method to lower carbon dioxide (CO 2 ) emissions is to switch to renewable energy sources. However, the transition to renewable energy such as solar and wind is likely to take decades, hence the continued reliance on non-sustainable energy sources [1,2]. In order to meet the short-to medium-term emissions targets, the capture, storage and utilization of CO 2 from large anthropogenic point sources such as fossil fuel power plants and the steel industry is key to mitigating the ever-increasing levels of CO 2 in the atmosphere, thereby preventing irreversible climate change [2][3][4].
One of the key challenges facing materials for carbon capture from anthropogenic point sources is the low concentration of CO 2 in flue gas streams-often around or lower than 20%. The remaining volume comprises largely N 2 with smaller amounts of water vapor, oxygen, SO 2 and NO x (among others) [5][6][7][8]. In order to capture CO 2 efficiently, any capture process therefore needs to show high selectivity toward CO 2 .
The current state-of-the-art industrial method of capturing CO 2 , dubbed amine scrubbing, has remained unchanged for decades and involves the use of aqueous solutions of amines such as monoethanolamine (MEA) [9]. This process relies on chemisorption, by which the MEA selectively reacts with CO 2 to form a carbamate salt.
Over the past few years, there has been a move away from the amine scrubbing process due to significant and numerous drawbacks. These include the chemisorption process requiring very high temperatures (ca. 130 • C) to liberate the CO 2 and regenerate the free amine. Attaining these high temperatures is a challenge for the industry and comes short of the PAF/PPN materials, HCP synthesis is considerably less complex, cheaper and more readily scalable.
Hypercrosslinked polymers are rigid porous networks with typical surface areas in the range of 500-2000 m 2 /g [64][65][66]. Their synthesis is often based on Friedel-Crafts chemistry using a Lewis acid catalyst such as iron (iii) chloride to yield a highly crosslinked and permanently microporous insoluble solid product. HCP synthesis requires the use of crosslinking groups, such as methyl chlorides, often dubbed "internal crosslinkers" [67,68], or external crosslinkers such as formaldehyde dimethyl acetal (FDA) [69]. This external "knitting method" allows potentially any rigid aromatic monomer to be hypercrosslinked.
Crucially, the "knitting method" provides a route for the incorporation of a range of chemical functionalities into the networks by polymerization of pre-functionalized monomers. This has led to the investigation of HCPs for a variety of different applications [27,[70][71][72]. For CO 2 capture, it is well known that different chemical moieties can impart increased selectivity toward CO 2 over other gases due to more favorable interactions with the chemical moiety and the dipole of the CO 2 [31,[73][74][75]. These interactions are crucial to maximizing their selectivity toward CO 2 .
In this work we report the synthesis, characterization and implementation of functional HCP networks for use as solid sorbents using a PSA approach. The CO 2 uptake capacity and uptake kinetics are measured at high pressure followed by measurements using simulated flue gas compositions. The CO 2 selectivity of the materials is calculated, and the recyclability potential of the HCPs is evaluated. Further to this, in order to keep the study industrially applicable, all samples were exposed to more industrially relevant gas streams, and the materials themselves were exposed to the humid laboratory conditions and not merely used immediately after drying.

Results and Discussion
Seven hypercrosslinked polymers were synthesized from functionalized monomers, all possessing different chemical moieties to study how these groups affected the CO 2 uptake and selectivity at high pressure. Monomers including alcohol functionalities (triphenylmethanol and 1,1 -bi-2-naphthol (BINOL)) and amine functionality (2 • amine carbazole and 3 • amine triphenylamine), which have been previously shown to increase CO 2 uptakes in porous polymers [24], were investigated. Other functionalities such as halogens (fluorobenzene) and a newly synthesized network based on dibenzyl ether were also investigated ( Figure 1). Further to this, a non-functionalized network was synthesized from polystyrene which provides a non-functionalized network, which could also be synthesized from waste polymers [76]. This network provides a control with which to compare the functionalized networks. While hypercrosslinked polymers made from poly(styrene) [77], carbazole [78], BINOL [27], triphenylamine [79] and fluorobenzene [70] have previously been reported, this is, to our knowledge, the first reported synthesis of networks synthesized from dibenzyl ether and triphenylmethanol.
All networks were obtained in good yields (Table S1) similar to previous reports of the same networks [27,69]. Structural characterization of the HCPs was performed by elemental analysis (Table S2), infrared spectroscopy (FT-IR) ( Figure S1) and 13 C solid state cross polarization magic-angle spinning (CP/MAS) NMR spectroscopy (ssNMR) (Figure 2 and Figure S2). Calculated %C, H and N of the networks were found to be typical for HCPs synthesized via Friedel-Crafts alkylation. There is some variation from the expected values as these are calculated assuming an idealized structure in which all protons have been exchanged for a methylene bridge. The presence of end groups and adsorbed molecules such as CO 2 and water vapor may also contribute to the deviation from theoretical values. Furthermore, the presence of residual iron known to persist even after extensive purification procedures may also lead to errors in these values. However, incorporation of heteroatoms such as nitrogen indicates successful polymerization, and nitrogen values of 5.17% and 4.35% were observed for the carbazole and triphenylamine networks, respectively, indicating successful incorporation of amines into the structure. All networks were obtained in good yields (Table S1) similar to previous reports of the same networks [27,69]. Structural characterization of the HCPs was performed by elemental analysis (Table S2), infrared spectroscopy (FT-IR) ( Figure S1) and 13 C solid state cross polarization magic-angle spinning (CP/MAS) NMR spectroscopy (ssNMR) ( Figures  2 and S2). Calculated %C, H and N of the networks were found to be typical for HCPs synthesized via Friedel-Crafts alkylation. There is some variation from the expected values as these are calculated assuming an idealized structure in which all protons have been exchanged for a methylene bridge. The presence of end groups and adsorbed molecules such as CO2 and water vapor may also contribute to the deviation from theoretical values. Furthermore, the presence of residual iron known to persist even after extensive purification procedures may also lead to errors in these values. However, incorporation of heteroatoms such as nitrogen indicates successful polymerization, and nitrogen values of 5.17% and 4.35% were observed for the carbazole and triphenylamine networks, respectively, indicating successful incorporation of amines into the structure.
Analysis by FTIR ( Figure S1) suggests that the incorporation of the monomers into the networks was successful with characteristic signals at ca. 2800 cm −1 and 1600 cm −1 corresponding to the C-H and C=C stretches for each network. Additional signals at ca. 3500 cm −1 and ca. 1000 cm −1 are assigned to the -OH and ether stretch of the triphenylmethanol and dibenzyl ether network, respectively. 13 C ssNMR spectra were collected for all samples and can be seen in Figure S2, while the spectra for the two newly synthesized materials are presented in Figure 2. All networks showed two prominent signals at ca. 140 and 130 ppm corresponding to quaternary aromatic carbons (CAr) and aromatic CAr-H. Signals at 36 ppm are assigned to methylene bridges in the networks. The resonance at 51 ppm for the triphenylmethanol network is assigned to the C-OH. For the dibenzyl ether network, a resonance at 72 ppm is assigned to the CH2-O-CH2 carbons adjacent to the ether linkage. A further resonance is observed at ca. 17 ppm and is attributed to unreacted end groups arising from incomplete reaction of the FDA crosslinker.
The porosity of the networks was measured using nitrogen adsorption/desorption isotherms at 77 K ( Figure 3). Brunauer-Emmett-Teller (BET) surface areas were calculated over a relative pressure range (P/P0) of 0.01-0.15 with the total pore and micropore volumes calculated at 0.95 and 0.1 P/P0, respectively (Table 1). All networks adsorbed large volumes of nitrogen at low relative pressure (<0.1 P/P0), indicating the presence of micropores. All networks demonstrated further uptake at higher partial pressures. This was particularly noticeable for Analysis by FTIR ( Figure S1) suggests that the incorporation of the monomers into the networks was successful with characteristic signals at ca. 2800 cm −1 and 1600 cm −1 corresponding to the C-H and C=C stretches for each network. Additional signals at ca. 3500 cm −1 and ca. 1000 cm −1 are assigned to the -OH and ether stretch of the triphenylmethanol and dibenzyl ether network, respectively. 13 C ssNMR spectra were collected for all samples and can be seen in Figure S2, while the spectra for the two newly synthesized materials are presented in Figure 2. All networks showed two prominent signals at ca. 140 and 130 ppm corresponding to quaternary aromatic carbons (C Ar ) and aromatic C Ar-H . Signals at 36 ppm are assigned to methylene bridges in the networks. The resonance at 51 ppm for the triphenylmethanol network is assigned to the C-OH. For the dibenzyl ether network, a resonance at 72 ppm is assigned to the CH 2 -O-CH 2 carbons adjacent to the ether linkage. A further resonance is observed at ca. 17 ppm and is attributed to unreacted end groups arising from incomplete reaction of the FDA crosslinker.
The porosity of the networks was measured using nitrogen adsorption/desorption isotherms at 77 K ( Figure 3). Brunauer-Emmett-Teller (BET) surface areas were calculated over a relative pressure range (P/P 0 ) of 0.01-0.15 with the total pore and micropore volumes calculated at 0.95 and 0.1 P/P 0 , respectively (Table 1). All networks adsorbed large volumes of nitrogen at low relative pressure (<0.1 P/P 0 ), indicating the presence of micropores. All networks demonstrated further uptake at higher partial pressures. This was particularly noticeable for the poly(styrene) network, which demonstrates a type II hysteresis loop on the desorb indicative of further larger (meso)pores, consistent with previously reports [77].
All samples were found to be porous with specific BET surface areas ranging from 213 m 2 /g to 1124 m 2 /g. The highest surface area was found to be derived from the polystyrene network and is similar to that reported previously in the literature [77]. Overall, the inclusion of functionality into the networks resulted in a lower surface area than nonfunctionalized HCPs. This is due to the functional groups sitting in the pores of the material and also the fact that different chemical moieties can hinder the efficacy of the hypercrosslinking process, thereby resulting in a less crosslinked network. Functional 3D monomers, however, such as BINOL, are still able to produce relatively high surface area networks. Despite their lower surface areas, the effects of the functionality are still interesting for CO 2 capture and the potential for increased selectivity over nitrogen.
inclusion of functionality into the networks resulted in a lower surface area than nonfunctionalized HCPs. This is due to the functional groups sitting in the pores of the material and also the fact that different chemical moieties can hinder the efficacy of the hypercrosslinking process, thereby resulting in a less crosslinked network. Functional 3D monomers, however, such as BINOL, are still able to produce relatively high surface area networks. Despite their lower surface areas, the effects of the functionality are still interesting for CO2 capture and the potential for increased selectivity over nitrogen.   inclusion of functionality into the networks resulted in a lower surface area than nonfunctionalized HCPs. This is due to the functional groups sitting in the pores of the material and also the fact that different chemical moieties can hinder the efficacy of the hypercrosslinking process, thereby resulting in a less crosslinked network. Functional 3D monomers, however, such as BINOL, are still able to produce relatively high surface area networks. Despite their lower surface areas, the effects of the functionality are still interesting for CO2 capture and the potential for increased selectivity over nitrogen.    The total pore volumes of the materials ranged from 0.14 cm 3 /g to 1.01 cm 3 /g, with the fluorobenzene-and polystyrene-derived HCPs showing the lowest and highest pore volumes, respectively, as might be expected from the highest and lowest surface area networks. This is also consistent with the pore size distribution plots ( Figure S3), which show that only micropores are present in the fluorobenzene network yet an abundance of mespores are micropores are present in the poly(styrene) network. As a proportion of pore volume (V 0.1 /V tot ), both carbazole and fluorobenzene showed the largest contribution of micropores, while dibenzyl ether was found to have a larger proportion of meso-and macropores. It has been previously reported that smaller pores are preferential over larger pores for CO 2 capture, particularly at lower pressures where the uptake has not reached a maximum. It was hypothesized that the networks with a larger % of micropores may therefore be better suited toward CO 2 capture than those possessing larger pores at 25 bar [28].

Kinetic Uptake of CO 2
High-pressure CO 2 adsorption experiments were conducted using the setup as previously reported by Reed and co-workers [23,80]. Briefly, an adsorbent was packed into a sealed tube, which was exposed to high pressures of gas before being weighed to gravimetrically determine the CO 2 uptake. All measurements were repeated three times and an average of the data was taken and used. All measurements on the functionalized HCPs were recorded at 40 • C to more closely match cooled flue gas temperatures from industrial sources. Stack temperatures can vary depending on the process but can range from 120 • C for post-combustion processes to 250-350 • C for steel plants and over 1000 • C for smelting works. As such, flue gas temperatures need to be reduced to values where absorption or adsorption are feasible. Moisture vapor is also an important consideration when for post-combustion capture [5,26]. All samples were therefore tested under "wet" conditions. More specifically, after synthesis, the samples were dried under vacuum at 60 • C to remove solvents before then being allowed to adsorb moisture from the air at 40-50% humidity for at least 24 h before any adsorption measurements were carried out. These conditions allow for results more comparable to those used in industry where gas mixtures are typically hydrated.
Pressures of 10 and 20 bar are typical pressures for PSA, which are easily attainable without a significant increase in plant operating costs. The rate at which each network reached saturation at 10 and 20 bar was therefore measured (Figure 4). At 20 bar, all HCP networks become fully saturated rapidly with t 90 values (the time at which 90% of the total uptake is completed) of 85 s or less (Table S5), while at the lower pressure of 10 bar, the time to reach saturation was up to 3 min, with the hydrophilic networks triphenylmethanol and BINOL taking longest and the hydrophobic networks poly(styrene) and fluorobenzene the shortest (Figure 4). The rapid sorption period is advantageous should these materials be applied to an industrial PSA approach given that the less time the material has to spend at elevated pressures, the greater the economic and energy benefit.
At 10 bar, the two -OH-containing networks (triphenylmethanol and BINOL) showed the highest final uptakes. Alcohol-containing porous polymers have previously been shown to demonstrate good CO 2 capture capabilities, and these results further corroborate the advantage of such functionalities even at higher pressures [27,81]. This is closely followed by the triphenylamine network, which demonstrates uptake of close to 15% wt. The highest surface area material-the non-functional poly(styrene)-shows uptake at 10 bar at around 11% wt. This material has a much higher surface area than the two alcohol materials yet underperforms in comparison to these two networks. Finally, the carbazole, fluorobenzene and the newly synthesized dibenzyl ether networks all demonstrate less prominent uptakes ranging from 6% wt. to 10% wt. At 10 bar, the two -OH-containing networks (triphenylmethanol and BINOL) showed the highest final uptakes. Alcohol-containing porous polymers have previously been shown to demonstrate good CO2 capture capabilities, and these results further corroborate the advantage of such functionalities even at higher pressures [27,81]. This is closely followed by the triphenylamine network, which demonstrates uptake of close to 15% wt. The highest surface area material-the non-functional poly(styrene)-shows uptake at 10 bar at around 11% wt. This material has a much higher surface area than the two alcohol materials yet underperforms in comparison to these two networks. Finally, the carbazole, fluorobenzene and the newly synthesized dibenzyl ether networks all demonstrate less prominent uptakes ranging from 6% wt. to 10% wt.
At 20 bar, all samples show increased uptake of CO2 compared to 10 bar, which is to be expected. The triphenylmethanol network continues to show the highest final uptake, being close to 35% wt. However, the largest improvement in uptake is exhibited by the poly(styrene) network, with an uptake of just over 25% wt. This result perhaps highlights that at elevated pressures, it is the surface area which becomes more important than chemical functionality. The BINOL network shows a final uptake of just under 20% wt., as does the triphenylamine network. The newly synthesized dibenzyl ether network shows the poorest uptake at ~10% wt., which is not much of an improvement from the uptake reported at 10 bar. This poor performance, despite a reasonable surface area, could be somewhat due to the presence of larger pore sizes dominating the material. In comparison, the triphenylmethanol, carbazole and fluorobenzene networks have a greater proportion of smaller micropores aiding their uptake under these conditions. Finally, the carbazole and fluorobenzene networks both have uptakes of ~15% wt., which is a slight increase from the values recorded at 10 bar.

Selectivity Measurements
In order to investigate how selective networks were for CO2 over that of the major component of flue gas (N2), individual uptakes of both CO2 and N2 were measured for each HCP network at pressures between 5 and 25 bar at a temperature of 40 °C ( Figures  S4 and S5). HCP networks were exposed to a pressurized stream of either pure CO2 or N2 for a 5 min adsorption period, the time at which the previous kinetic runs were shown to be sufficient for equilibration, after which the gravimetric uptake was recorded, and the average uptake was calculated over three runs. Using these experiments, it is possible to estimate the CO2:N2 selectivity of the networks at high pressures typical for PSA.
All the networks demonstrated much higher uptakes of CO2 than N2 under identical adsorption conditions (Table 2), typically thereby demonstrating a preference toward adsorbing CO2 over N2. The nitrogen uptake of the networks correlates well with the BET surface areas of the materials, exhibiting no strong interaction with the network surface At 20 bar, all samples show increased uptake of CO 2 compared to 10 bar, which is to be expected. The triphenylmethanol network continues to show the highest final uptake, being close to 35% wt. However, the largest improvement in uptake is exhibited by the poly(styrene) network, with an uptake of just over 25% wt. This result perhaps highlights that at elevated pressures, it is the surface area which becomes more important than chemical functionality. The BINOL network shows a final uptake of just under 20% wt., as does the triphenylamine network. The newly synthesized dibenzyl ether network shows the poorest uptake at~10% wt., which is not much of an improvement from the uptake reported at 10 bar. This poor performance, despite a reasonable surface area, could be somewhat due to the presence of larger pore sizes dominating the material. In comparison, the triphenylmethanol, carbazole and fluorobenzene networks have a greater proportion of smaller micropores aiding their uptake under these conditions. Finally, the carbazole and fluorobenzene networks both have uptakes of~15% wt., which is a slight increase from the values recorded at 10 bar.

Selectivity Measurements
In order to investigate how selective networks were for CO 2 over that of the major component of flue gas (N 2 ), individual uptakes of both CO 2 and N 2 were measured for each HCP network at pressures between 5 and 25 bar at a temperature of 40 • C ( Figures S4  and S5). HCP networks were exposed to a pressurized stream of either pure CO 2 or N 2 for a 5 min adsorption period, the time at which the previous kinetic runs were shown to be sufficient for equilibration, after which the gravimetric uptake was recorded, and the average uptake was calculated over three runs. Using these experiments, it is possible to estimate the CO 2 :N 2 selectivity of the networks at high pressures typical for PSA.
All the networks demonstrated much higher uptakes of CO 2 than N 2 under identical adsorption conditions (Table 2), typically thereby demonstrating a preference toward adsorbing CO 2 over N 2 . The nitrogen uptake of the networks correlates well with the BET surface areas of the materials, exhibiting no strong interaction with the network surface functionalities. By dividing the CO 2 uptake by the N 2 uptake, crude selectivity ratios can be calculated for each material, which serve as a rough guide as to the relative affinity the networks have for each gas. Perhaps unsurprisingly, the selectivity of each material decreases as pressure increases. As an example, the triphenylamine network has a selectivity of 9:1 at 10 bar, which decreases to 7:1 at 20 bar. Likewise, this is also true for the triphenylmethanol network, which has a calculated selectivity of 7:1 at 10 bar and 5:1 at 20 bar. The poly(styene) network, non-functional in nature, has the best selectivity of 13:1 and 9:1 at 10 bar and 0 bar, respectively. This selectivity is clearly not a result of chemical functionality and is most likely the result of the high surface area of the material and small pores, which preferentially adsorb CO 2 over N 2 . The dibenzyl ether and fluorobenzene networks both actually report slightly increased selectivity at higher pressures, though this is a result of the poor uptakes of both gases at these pressures. While some insight into the selectivity of the materials can be derived using pure gas streams, the use of mixed gas streams is more representative of actual industrial flue gas. To investigate how the materials performed at enriching a CO 2 stream, the most promising materials were exposed to a gas mix composed of 80:20 N 2 :CO 2 at 40 • C and 20 bar for 5 min. The concentration of CO 2 in the output gas at 20 bar was measured using an IR sensor, after which the pressure was then released from the adsorber. When the pressure reached 1 bar, the concentration of CO 2 was measured again. Finally, the same experiment was repeated using a stream composed of a 50:50 N 2 :CO 2 mix at 40 • C ( Figure 5). This test replicated two cycles, whereby the output from the first cycle is fed back into the PSA setup and the method is repeated once again.
Initially, when the chosen samples were exposed to an 80:20 N 2 :CO 2 stream, all materials tested were able to selectively adsorb the CO 2 at 20 bar and then desorb it at 1 bar. This resulted in the successful separation of CO 2 from N 2 and the generation of a gas stream enriched to over 50% CO 2 in the case of triphenylmethanol, poly(styrene) and carbazole after one cycle. We therefore exposed the materials to a 50:50 N 2 :CO 2 stream, the equivalent of feeding the stream from the first test back into the materials and repeated the experiment again. The triphenylmethanol and triphenylamine samples were able to enrich the stream of gas to over 80% CO 2 . This experiment demonstrates that these materials are able to take a dilute stream of flue gas and, after two pressure swing cycles, convert this dilute stream into a concentrated CO 2 stream by preferential adsorption of CO 2 over N 2 . Finally, the ability of the sorbents to be used over repeat adsorption-desorption cycles was tested using the best-performing triphenylmethanol and triphenylamine networks ( Figure 6). These networks were exposed to a 25-bar stream of CO2 before having the pressure reduced to 1 bar with the uptakes at each pressure recorded and repeating Finally, the ability of the sorbents to be used over repeat adsorption-desorption cycles was tested using the best-performing triphenylmethanol and triphenylamine networks ( Figure 6). These networks were exposed to a 25-bar stream of CO 2 before having the pressure reduced to 1 bar with the uptakes at each pressure recorded and repeating for 10 cycles. Importantly, as in a typical PSA process, the materials were not exposed to a vacuum between runs to remove any CO 2 so as not to further increase the energy demands of the process. Both materials reached a maximum uptake at 25 bar, and this was found to be reproducible over the 10-cycle run, demonstrating no loss in performance over time. Both samples retained some gas at 1 bar, though this quantity was minimal (<3% wt.) and had no significant effect on the uptake at higher pressures. Finally, the ability of the sorbents to be used over repeat adsorption-desorption cycles was tested using the best-performing triphenylmethanol and triphenylamine networks ( Figure 6). These networks were exposed to a 25-bar stream of CO2 before having the pressure reduced to 1 bar with the uptakes at each pressure recorded and repeating for 10 cycles. Importantly, as in a typical PSA process, the materials were not exposed to a vacuum between runs to remove any CO2 so as not to further increase the energy demands of the process. Both materials reached a maximum uptake at 25 bar, and this was found to be reproducible over the 10-cycle run, demonstrating no loss in performance over time. Both samples retained some gas at 1 bar, though this quantity was minimal (<3% wt.) and had no significant effect on the uptake at higher pressures.

Conclusions
To conclude, a series of functional porous materials synthesized via conventional hypercrosslinking chemistry were applied as sorbents to selectively separate CO2 from simulated flue gas mixes. The uptake capacity, uptake rate and CO2:N2 selectivity at high pressure were all thoroughly examined to test the materials at high pressure and using a pressure swing approach. The best-performing materials were then taken forward and

Conclusions
To conclude, a series of functional porous materials synthesized via conventional hypercrosslinking chemistry were applied as sorbents to selectively separate CO 2 from simulated flue gas mixes. The uptake capacity, uptake rate and CO 2 :N 2 selectivity at high pressure were all thoroughly examined to test the materials at high pressure and using a pressure swing approach. The best-performing materials were then taken forward and applied toward actual pressure swing separation experiments using simulated gas mixtures representative of those in industry. Finally, the recyclability of the optimum materials was tested to investigate if their performance was hindered after multiple adsorb/desorb cycles. All materials were found to uptake CO 2 rapidly, with most of the uptake being complete within 2 min with the -OH functionalized and non-functional poly(styrene) network showing the highest CO 2 capacity. Due to their high and selective uptakes, both the triphenylmethanol and triphenylamine networks were taken forward and applied to an actual pressure swing approach, where it was found that after only two cycles, they were able to convert a 20% CO 2 stream into one exceeding 85% CO 2 . This was an excellent example of how cheaply synthesized porous materials can be easily synthesized and applied to a pressure swing methodology, demonstrating excellent CO 2 :N 2 capabilities. It is hoped that this work inspires more research into PSA techniques to improve on the current energy-intensive and fiscally demanding temperature swing techniques rife throughout the industry.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in the supplementary information.