High-Performance Porous Organic Polymers for Environmental Remediation of Toxic Gases

Sulfur dioxide (SO2) is a harmful acidic gas generated from power plants and fossil fuel combustion and represents a significant health risk and threat to the environment. Benzimidazole-linked polymers (BILPs) have emerged as a promising class of porous solid adsorbents for toxic gases because of their chemical and thermal stability as well as the chemical nature of the imidazole moiety. The performance of BILPs in SO2 capture was examined by synergistic experimental and theoretical studies. BILPs exhibit a significantly high SO2 uptake of up to 8.5 mmol g–1 at 298 K and 1.0 bar. The density functional theory (DFT) calculations predict that this high SO2 uptake is due to the dipole–dipole interactions between SO2 and the functionalized polymer frames through O2S(δ+)···N(δ–)-imine and O=S=O(δ–)···H(δ+)-aryl and intermolecular attraction between SO2 molecules (O=S=O(δ–)···S(δ+)O2). Moderate isosteric heats of adsorption (Qst ≈ 38 kJ mol–1) obtained from experimental SO2 uptake studies are well supported by the DFT calculations (≈40 kJ mol–1), which suggests physisorption processes enabling rapid adsorbent regeneration for reuse. Repeated adsorption experiments with almost identical SO2 uptake confirm the easy regeneration and robustness of BILPs. Moreover, BILPs possess very high SO2 adsorption selectivity at low concentration over carbon dioxide (CO2), methane (CH4), and nitrogen (N2): SO2/CO2, 19–24; SO2/CH4, 118–113; SO2/N2, 600–674. This study highlights the potential of BILPs in the desulfurization of flue gas or other gas mixtures through capturing trace levels of SO2.


■ INTRODUCTION
The emission of acidic gases from fossil fuel consumption is a serious threat to human health and the environment. 1,2Apart from carbon dioxide's (CO 2 ) greenhouse effect, which contributes to global warming and climate change, the effects of sulfur dioxide (SO 2 ) on the environment are even more hazardous.Particularly, the acid rain caused by SO 2 is responsible for deforestation and threatens animal life cycles. 3ccording to the National Library of Medicine, the exposure of SO 2 over 100 ppm in various species is deadly, 4 and its concentration should not exceed 0.5 ppm more than once per year for 3 h according to the US EPA. 5 Therefore, it is highly desirable to develop efficient materials that can selectively remove SO 2 from the source gas mixtures.6][7][8][9]11 However, sulfate byproducts, acidic wastewater, and solvent volatilization and recycling at high temperatures represent serious concerns that still need to be addressed. Th use of porous solid adsorbents in capturing toxic gases is a promising method and has been extensively studied over the last two decades.In recent years, there has been a significant interest in studying the SO 2 capture by synthesized porous materials both purely organic structures and organic−inorganic hybrid structures commonly known as metal−organic frameworks (MOFs).12−19 In particular, MOF-177 that has an ultrahigh specific surface area (4100 m 2 g −1 ) possessed a remarkable SO 2 uptake of 25.7 mmol g −1 at 293 K and 1.0 bar.20 Similarly, NOTT-202a (2220 m 2 g −1 ) exhibited 13.6 mmol g −1 at 268 K. 14 Although MOFs are promising because of their high surface areas and highly ordered structures, the acidic nature of SO 2 particularly in the presence of moisture may destabilize the frameworks.In that aspect, porous organic polymers that are linked with strong covalent bonds appear to be more promising in environmental remediation.21−23 A recent report on benzimidazole-derived carbons (BIDCs) showed very high SO 2 sorption capacity up to 21.42 mmol g −1 at 298 K and 1 bar.12 BIDCs were made by the pyrolysis of imidazole in the presence of KOH at high temperatures.BIDC-3-800, which was made by pyrolysis at 800 °C, is highly porous with a surface area of 3750 m 2 g −1 and exhibited the best performance of SO 2 uptake (21.42 mmol g −1 at 298 K and 1 bar), while BIDC-2-700 prepared at 700 °C (surface area 1200 m 2 g −1 ) showed SO 2 uptake of 10.25 mmol g −1 at the same condition.High surface area and heteroatom functionality are expected to play important roles in high SO 2 uptake.Functionalized polymers of intrinsic microporosity (PIMs) also showed high SO 2 uptake (7.32 mmol g −1 at 298 K and 1 bar).24 The recent report on SO 2 capture by cage molecules showed that the tertiary amine has higher SO 2 capture capabilities than the secondary amine or imine functionalities.16 SO 2 has a very high affinity toward primary amines, and it can be used as the connector between two primary amine functional groups through the formation of sulfonamide linkage.25 These findings encouraged us to study the effect of heteroatoms, particularly imidazole rings, on the selective SO 2 capture. Ov28 BILPs also work as promising metal-free photocatalysts for CO 2 reduction.29 Structural features such as high surface area, pore functionality with heteroatoms, and physicochemical stability make BILPs ideal for acidic SO 2 capture and separation.
Herein, we studied the effects of the chemical and textural properties of BILPs on SO 2 capture.Both experimental and theoretical studies have been performed for two representative BILPs, namely, BILP-3 and BILP-4 (Figure 1).We demonstrate that BILPs are well-suited for selective SO 2 capture because of their high thermal and chemical stability, permanent porosity, and exceptionally high uptake capacity, which are among the highest for all known covalently linked porous organic materials such as COFs, POPs, and cage molecules reported to date.Our experimental and theoretical studies reveal that the electronic nature of the benzimidazole moiety in the polymer frame plays a key role in selective and high SO 2 uptake.

■ EXPERIMENTAL SECTION
Materials and Methods.All chemicals, including the monomer 1,2,4,5-benzenetetramine (BTA) used in this research, were purchased from commercial suppliers (Sigma-Aldrich, Acros Organ-ics, or Frontier Scientific) and used as received.Synthesis and characterization of other monomers (tetrakis(4-formylphenyl)methane (TFPM) and 2,3,6,7,14,15-hexaaminotriptycene (HATT)) and polymers BILP-3 and BILP-4 were reported in our earlier publications. 26,27Synthesized polymers were characterized by performing thermogravimetric analysis (TGA), Fourier-transform infrared spectroscopy (FT-IR), and X-ray photoelectron spectroscopy (XPS) measurements.The details of these measurements are discussed in Supporting Information.Surface areas and porosity were calculated from nitrogen adsorption isotherms collected using a Quantachrome Autosorb 1-C volumetric analyzer.The SO 2 uptakes were studied using the facilities at Northwestern University (NU) and Pacific Northwest National Laboratory (PNNL).For volumetric equilibrium SO 2 uptakes measured at NU, the materials were tested by using pure SO 2 gas on a Micromeritics 3Flex Surface Analyzer at 298 K equipped with a corrosive resistance-enhanced manifold.The 3Flex was set up in a walk-in fume hood, and the environment was monitored with a SO 2 sensor.Samples were degassed at 60 °C overnight before any gas adsorption experiments.The recyclability of the material was tested by reanalyzing the material for each toxic gas a second time after activation under vacuum at room temperature to remove adsorbed gas from the previous analysis.The dynamic sorption system used at the PNNL to measure SO 2 adsorption and desorption isotherms is shown in Figure S1 and is discussed there.
Computational Methodology.To reduce the computational demand, polymeric BILPs were modeled by terminating the corresponding segment of the benzimidazole-containing units with hydrogen atoms.The segment used for BILP-3 calculation has 3 benzimidazole units, while the segment used for BILP-4 calculation has 2 benzimidazole units (Figure S6).The binding affinities (BE) per SO 2 of BILP-3 and BILP-4 were calculated using the following equation: Here, BILPs@nSO 2 represents the BILPs interacting with nSO 2 molecules (n is the number of SO 2 molecules).E[BILPs@nSO 2 ] is the total energy of BILPs@nSO 2 , while E[BILPs] is the energy of the polymer without SO 2 and E[nSO 2 ] is the total energy of nSO 2 molecules.The number of SO 2 molecules (n) is chosen based on the number of nitrogen centers within the polymer segment used in the calculation.For BILP-3, n values are 3 (half of the total nitrogen centers) and 6 (equal to the total nitrogen centers).Similarly, the number of SO 2 molecules (n) for BILP-4 is 2 (half of the total nitrogen centers) and 4 (equal to the total nitrogen centers).The density functional theory (DFT) calculations were performed using two forms of exchange−correlation potentials: local density approximation functional (LDA) 30 and hybrid meta exchange− correlation functional, M06. 31 The M06 functional is particularly important because it accounts for the dispersive forces, which may play a key role in the case of weak interaction between the BILP substrate and SO 2 molecules.The LDA, on the other hand, overestimates binding 32 and, in some cases, yields binding energies closer to experiments when interactions are weak.However, this agreement results from the fortunate cancellation of errors as LDA does not include long-range dispersive forces.All calculations were carried out using Gaussian 09 33 software and 6-311+G* 34 basis sets.
In obtaining the equilibrium geometries, the SO 2 molecules were allowed to approach the top and bridge sites of the central ring as well as the top and planar sites of N atoms.The SO 2 molecule was also aligned horizontally and vertically to the plane of the benzimidazole moiety.Geometry optimization was initially performed with LDA and then further with M06.Minimum in the potential energy surface (PES) was achieved to ensure the most stable geometry.Natural bonding orbital method (NBO) 35 was used to calculate the atomic charges.

■ RESULTS AND DISCUSSION
The synthesis of BILP-3 and BILP-4 and their porosities were reproduced according to our published methods (Scheme 1). 26,27Briefly, condensation reactions between aldehyde and HCl salt form of amine monomers (TFPM and HATT, or BTA) afforded BILP-3 and BILP-4, respectively.The solids were collected as yellow powders in good yields.After washing with aqueous 2 M HCl, 2 M NaOH, deionized water, and acetone, the solid products were dried at first at ambient conditions and then under reduced pressure at 120 °C.
Nitrogen sorption isotherms were collected for the activated samples at 77 K.The calculated Brunauer−Emmett−Teller (BET) surface area and pore size distribution using nonlocaldensity function theory (NLDFT) agreed well with the reported values: BILP-3 (SA: 1390 m 2 g −1 , PSD: 10.6 Å) and BILP-4 (SA: 1220 m 2 g −1 , PSD: 10.6 Å).It is worth noting that both BILPs have high chemical and thermal stabilities up to 400 °C based on reported TGA studies.
Volumetric SO 2 adsorption isotherms for BILP-3 were collected at 288, 298, and 308 K using the facility at Northwestern University and are shown in Figure 2A.Adsorption isotherms were also collected using the dynamic adsorption system for both BILP-3 and BILP-4 at 298 K using the facility at Pacific Northwest National Laboratory, and they are shown in Figure S2A.It has been found that there is a slight deviation in the SO 2 uptake for the two measurements from each institution.We assume this slight difference is due to the difference in internal structures of two BILP-3 samples that were synthesized in two different batches independently at VCU and UW-Platteville.The BILPs are amorphous because of the irreversible nature of imidazole ring formation, and it is not surprising to obtain the final products with slightly

Langmuir
different internal structures from two independently prepared samples.The BET surface areas and gas uptakes sometimes vary significantly, even for well-defined crystalline substances.For example, based on the synthetic methods, MOF-5 possesses a BET surface area in a wider range from 260 to 4400 m 2 g −1 . 36ILPs show a type-I isotherm, and desorption follows a moderate degree of hysteresis.A similar hysteresis pattern was observed for the cage molecule 6FT-RCC3, which has been attributed to the flexibility of polymers. 16,37In contrast to cage molecules, BILPs show no open loop hysteresis, which occurs due to swelling effects. 16,38The SO 2 uptake of BILP-3 is 8.6 mmol g −1 (35 wt %) at 298 K and 1 bar as shown in Figure 2A.BILP-4 also shows a similar high SO 2 uptake of 6.1 mmol g −1 (28 wt %) at the same temperature and pressure (Figure S2A). 39The sharp initial uptakes at low pressure are consistent with the expected high affinity of the imidazole moieties for SO 2 .At ambient pressure, the uptakes by BILPs exceed most porous materials of similar surface areas investigated to date (Table S1); FMOF-2 (2.2 mmol g −1 at 298 K), 40 CoCo (2.5 mmol g −1 at 298 K), 41 MFM-520 (3.38 mmol g −1 at 298 K), 42 Pd(II)-based metal−organic case (6.0 mmol g −1 ), 17,43 and rival the best-performing MOFs: NOTT-202a (∼8 mmol g −1 at 293 K), NOTT-300 (8.1 mmol g −1 at 273 K), 44 and FM-300-(In) (8.28 mmol g −1 at 298 K) 45 even though some of these MOFs have much higher surface area (NOTT-202; SA BET = 2220 m 2 g −1 ). 14,40The highest SO 2 uptake of 25.7 mmol g −1 at 293 K and 1.0 bar was reported for MOF-177 (4100 m 2 g −1 ). 20The SO 2 uptake by BILP-3 at 298 K and 1 bar is comparable to TAM-POF (9.45 mmol g −1 ) 46 and Viologen-POF (11.3 mmol g −1 ) 47 and higher than the functionalized PIMs (PIM-1-AX: 7.32 mmol g −1 , PIM-1: 5.89 mmol g −1 , and PIM-1-COOH: 5.53 mmol g −1 ) 24 and sPANs (sPAN-1: 5.56, sPAN-1: 5.64 mmol g −1 ). 23The storage capacity of BILPs also surpasses many of the organic solvents and ionic liquids 9 and cage molecule CC3 (2.78 mmol g −1 ). 16The uptake by BILPs is somewhat lower than that of the KOH-treated benzimidazolederived carbons such as BIDC-700 (10.25 mmol g −1 ) and BIDC-800 with (21.42 mmol g −1 ) 12 and the best-performing cage molecule 6FT-RCC3 (13.78 mmol g −1 ). 16It is worth noting that BILP-3 possesses a significant SO 2 uptake (4.5 mmol g −1 ) at low pressure (0.1 bar) relevant for removing SO 2 trace contaminants from gas mixtures.This uptake is lower than the best-performing cage molecule 6FT-RCC3 (8.67 mmol g −1 ), 16 MOF MIL-125(Ti)-NH 2 (7.9 mmol g −1 ), 20 and COF CTF-CSU41 (6.7 mmol g −1 at 0.15 bar) 48 but higher than many other porous networks: MFM-170 (≈6.5 mmol g −1 ); 49 MOF-177 (1.0 mmol g −1 ); 20 CTF-CSU38 (4.4 mmol g −1 at 0.15 bar); 48 P(Ph-4MVIm-Br) (4.14 mmol g −1 ). 50o understand the high SO 2 uptake, the heat of adsorption (Q st ) at zero coverage for BILP-3 has been calculated from experimental SO 2 isotherms collected at 288, 298, and 308 K.The zero coverage Q st is 32.0 kJ mol −1 , which increases with SO 2 uptake to a maximum of 38.3 and then drops to 25.0 kJ mol −1 at 1.0 bar (Figure 2C).This value is similar to that reported for SO 2 interaction with NOTT-202a (35 kJ mol −1 ) 14 and viologen-POF (38.3 kJ mol −1 ). 47The Q st at zero coverage is lower than the Q st value reported for the cage molecule with a secondary amine (RCC3: 82.78 kJ mol −1 ) but similar to those case molecules with a tertiary amine (6FT-RCC3: 43.03 kJ mol −1 ) and imine nitrogen (38.46 kJ mol −1 ). 16The high Q st value for case molecule RCC3 suggests an almost irreversible chemisorption process, 16 while the moderate Q st for BILP-3 in this work suggests that it has a preference to undergo a physisorption-type interaction with SO 2 .Similar heat of adsorption values (29.2 and 32.3 kJ mol −1 ) were observed for cyanide (CN − ) containing ionic liquids [N(CN) 2 ] − -SO 2 51 and [C(CN) 3 ] − -SO 2 , respectively. 52Significantly higher heat of enthalpy was reported for ionic liquids that contain thiocyanate (SCN − ) anions: [SCN] − -SO 2 (73.0 kJ mol −1 ), which is due to chemisorption of SO 2 . 52SO 2 interacts chemically with the metal centers in some MOFs that also possess very high Q st values, for example, Mg-MOF-74 (>90 kJ mol −1 ) and Zn-MOF-74 (>70 kJ mol −1 ). 53Moderate Q st (35.83 and 28.39 kJ mol −1 ) has been reported for MOF-Th-Co-67 and MOF-Th-Co-66, respectively. 54o understand the reusability and stability of polymers upon exposure to SO 2 , BILPs were regenerated both at room temperature and at elevated temperature.For the regeneration processes (done at Northwestern University), BILP-3 was evacuated under vacuum at room temperature for 2 h prior to each adsorption data collection up to 1.0 bar.SO 2 uptake for BILP-3 was recorded for three cycles, as shown in Figure 2B.The SO 2 uptake slightly drops from the first cycle to the second cycle.However, it does not drop at all in the third cycle rather than slightly increases (BILP-3: 8.60, 8.12, and 8.29 mmol g −1 ).Almost identical SO 2 uptakes suggest that BILPs can be regenerated at room temperature and are reusable.Repeated uptake measurements were also done at the PNNL for both BILP-3 and BILP-4.BILPs were activated at an elevated temperature (150 °C under vacuum), and the volumetric adsorption experiments were repeated for activated samples.The isotherms for three cycles for both BILPs are depicted in Figure S2B, which confirm the same uptake tendency (BILP-3: 7.23, 6.63, and 6.64 mmol g −1 ; BILP-4: 6.04, 5.87, and 6.10 mmol g −1 ).The slight fluctuation in the SO 2 uptake during cycling is probably due to the flexible nature of the polymers, as stated earlier.A similar SO 2 uptake pattern was reported for viologen-POF, 47 BIDC, 12 TAM-POF, 46 and nanofiber aerogel. 55It is noteworthy to mention that BILPs are very robust under acidic and basic conditions.The synthesized BILPs were washed with 2.0 M HCl and NaOH.Therefore, we conclude that BILPs do not undergo any degradation or permanent chemical change under exposure to acidic SO 2 .The XPS spectra (discussed below) also preclude the chemisorption of SO 2 by BILPs.These repeated SO 2 uptake measurements thus suggest that BILPs are stable in SO 2 environment and, can be regenerated at room temperature with the same efficiency as at elevated temperatures and can be reused effectively in SO 2 capture.This result is very promising in terms of real-life applications of BILPs in SO 2 capture compared to ionic liquids, which require high temperatures to regenerate the adsorbents. 9,56he chemical and thermal stability of BILPs after exposure to SO 2 was also confirmed by TGA traces (Figure S3), which demonstrate two distinct weight losses.The initial weight loss at room temperature is due to the desorption of weak physisorbed SO 2 , while the second step started at ca. 200 °C and is proposed to be due to the desorption of strong physisorbed SO 2 .The strong physisorption is supported by the observation of a gradual weight loss upon heating to high temperatures.The total weight loss up to 300 °C is ∼30 wt %, which matches well with the SO 2 uptake obtained from the adsorption isotherms.No significant weight loss was observed except desorption of the above calculated amount of adsorbed SO 2 , which demonstrates the stability of BILPs under SO 2 Langmuir adsorption environments.Although desorption of strongly physisorbed SO 2 required high temperatures, it could be facilitated at lower temperatures under a vacuum, as observed in repeated SO 2 adsorption−desorption experiments.
Spectroscopic Characterization.The presence of adsorbed SO 2 on BILPs was examined by FT-IR and XPS spectroscopy.The FT-IR spectra were recorded for BILPs before and after exposure of samples to SO 2 (Figure S4).New sharp infrared (IR) peaks were observed at 1250 and 1120 cm −1 for SO 2 -dosed samples (BILPs@nSO 2 ) and were assigned to asymmetric and symmetric stretching vibrational modes of adsorbed SO 2 , respectively.These two vibrational modes for gas phase free SO 2 are observed at 1351 and 1147 cm −1 , respectively. 57,58The lower frequency shifts of adsorbed SO 2 compared to free SO 2 arise from strong interactions of SO 2 molecules with BILPs.It should be noted that larger downshifts, 1230 and 957 cm −1 , were observed for chemically adsorbed SO 2 in ionic liquids through sulfate formation (S�O and S−O stretches, respectively). 59An intense band appearing at 617 cm −1 is assigned to the bending mode of adsorbed SO 2 .A similar bending mode (624 cm −1 ) was reported for adsorbed SO 2 onto tertiary amine-based Merrifield resins. 60The significantly higher frequency shift of the bending mode compared to isolated gas phase SO 2 (508−518 cm −1 ) 61 was explained by a noncovalent charge transfer between SO 2 and the tertiary amine group. 60PS measurements were also used to explore the presence of adsorbed SO 2 on the BILPs.Figure S5 shows the survey XPS spectra for BILPs before and after exposure to SO 2 .Detailed regional scans and deconvoluted peaks of S 2p, N 1s, and O 1s are shown in Figure 3.The intense C 1s peak at 284.1 eV in each spectrum (Figure S5) confirms the presence of aromatic carbons (sp 2 carbon). 62−66 The presence of the O 1s at around 532 eV in BILPs (Figure 3d) before exposure to SO 2 most likely originates from unreacted aldehyde functional groups on the BILP surface.However, there is a significant intensity increase of the O 1s peak in BILPs@nSO 2 compared to BILPs.This indicates significant adsorption of SO 2 onto BILPs.The doublet pattern (398.6 and 400.4 eV) of the N 1s peak in BILPs (Figure 3a,b) with almost equal intensities was observed, which is consistent with the presence of two nitrogen sites with two different chemical environments in benzimidazole moieties: one is imine-type double-bonded nitrogen and the other is singlebonded nitrogen called pyridinic and pyrrolic nitrogen, respectively. 63Upon exposure to SO 2 , these peaks shift slightly Figure 3. High-resolution XPS spectra for BILPs and BILPs@nSO 2 : (a) N 1s for BILP-3 and BILP-3@nSO 2 ; (b) N 1s for BILP-4 and BILP-4@ nSO 2 ; (c) S 2p 3/2 and S 2p 1/2 for BILP-3@nSO 2 and BILP-4@nSO 2 ; (d) O 1s for BILP-3, BILP-4, BILP-3@nSO 2 , and BILP-4@nSO 2 .
to higher energy (398.7 and 400.9 eV), but the relative peak intensity of pyridinic nitrogen significantly decreases compared to that of pyrrolic nitrogen.This intensity alteration indicates that the pyridinic nitrogen in the benzimidazole ring is more affected by the adsorbed SO 2 .New peaks at around 169 and 232 eV were observed, which are assigned to S 2p and S 2s, respectively.−69 The presence of the S 2p 3/2 peak at 168.6 eV (Figure 3c) confirms the physisorption of SO 2 on BILPs.It has been reported that for multilayer physisorbed SO 2 , the S 2p 3/2 values are within 167.6−169.4−74 Theoretical Calculations.To understand the interaction between SO 2 and BILPs, we carried out calculations based on DFT methods.Figure S6 shows the electrostatic potential surface of optimized structures for the benzimidazolecontaining moieties of BILPs and SO 2 .As expected, the electronegative regions appear around the N-imine sites rather than those of N−H.A significant amount of electronegative charge is observed around the aromatic cores that appears to enhance the site-selective adsorption of polar SO 2 gas molecules on benzimidazole-derived surfaces.
Two geometries were taken into consideration for BILPs@ nSO 2 .One geometry considers a single SO 2 molecule per imidazole ring (n = 1), and the other considers two SO 2 molecules per imidazole ring (n = 2).For simplification, calculations were performed for the segments of BILP-3 and BILP-4 with three and two imidazole rings, respectively.In BILP-3, imidazole rings are well separated by the triptycene moiety, while in BILP-4, two imidazole rings are relatively closely separated by the benzene ring.Calculated results on BILP-4 are discussed first and then compared to BILP-3.Fully optimized geometries for BILP-4@nSO 2 using M06/6-311+G* are shown in Figure 4A,B.SO 2 molecules preferably lie close to the N-imine site when two SO 2 molecules were allowed to interact with BILP-4, with the distance between the N-imine site and SO 2 molecule being about 2.58 Å (d SN ) (Figure 4A).On the other hand, when four SO 2 molecules interact, two of them (called SO 2 (I)) lie close to the N-imine site with a distance of d SN = 2.49 Å, whereas the other two SO 2 molecules (called SO 2 (II)) lie in between aryl-H and N−H of imidazole rings (Figure 4B).To understand the bonding preference of SO 2 for N-imine and N−H of the imidazole ring, we examined the NBO charges on the interacting atoms.The NBO charges on the S atom (+1.62e) of SO 2 and N-imine (−0.56e) of the imidazole ring indicate the presence of a strong dipole−dipole interaction.The NBO charges on O atom (−0.86e) of SO 2 and the atom −H (+0.44e) of N−H indicate a relatively weaker hydrogen bonding.This suggests the preferential adsorption of SO 2 on N-imine of the imidazole ring.
The orientation of SO 2 (II) is such that the S atom is directed away from the H atom of the six-membered aryl ring (d SH = 3.13 Å) and is closer to the O atom of SO 2 (I) (d OS = 2.76 Å).On the other hand, one of the O atoms of SO 2 (II) is directed toward the H atom of N−H (d OH = 2.33 Å), while its distance from the H atom of aryl-H is 2.48 Å.This orientation of SO 2 (II) suggests an intermolecular attraction between the O atom of SO 2 (I) and the S atom of SO 2 (II) that slightly displaces the SO 2 molecules from the BILP-plane (Figure S7).Similar SO 2 −SO 2 interactions with an approximate intermolecular distance of 3.13 Å have been reported for MOF MFM-250. 42Geometry optimization was also performed using LDA/ Figure 4. Optimized geometries of BILPs@nSO 2 by using M06/6-311+G*.(A) BILP-4@2SO 2 , (B) BILP-4@4SO 2 , (C) BILP-3@3SO 2 , and (D) BILP-3@6SO 2 .BILP-4@nSO 2 was fully optimized, while BILP-3@nSO 2 was partially optimized for a constrained structure.Carbon is gray, nitrogen is blue, oxygen is red, sulfur is yellow, and hydrogen is white.The unit of bond length is angstrom (Å).
6-311+G*, and the optimized geometries of BILP-4@nSO 2 (Figure S8) showed a similar spatial orientation of SO 2 molecules, though the LDA method showed closer proximity to the adsorption sites.This is because the LDA method 32 overestimates binding.
Inspection of the electrostatic potential energy surface, as shown in Figure S6, revealed that the six-membered aryl ring possesses a significant electronegative charge that creates a partial positive charge on the attached hydrogen (aryl-H(δ + )).This may attract a partially negative O atom of SO 2 (II).To understand whether the aryl-H(δ + ) of six-membered rings or the N−H(δ+) of the imidazole ring has more attraction to (δ − )O�S�O(II), calculations were performed for BILP-3 in which the imidazole moieties are far apart, allowing a better understanding of the above effects.
The geometries of BILP-3@nSO 2 were optimized using both the LDA and M06 level of theory for a segment of polymer that contains the triptycene core.It should be noted that the interaction of six SO 2 molecules with BILP-3 results in bending or distortion of the BILP-3 strand when the geometry is fully optimized by using either LDA or M06.The distortion is relatively large for the LDA structure (Figure S9).Although this distortion for an isolated segment in the presence of SO 2 molecules may be explained in terms of strong intermolecular forces, such a large distortion in a real extended polymer network is unlikely.Therefore, a partial optimization was performed for a constrained structure of BILP-3@nSO 2 in which the polymer segment was frozen, and the binding sites of SO 2 molecules were only optimized (Figures S10 and S11). Figure 4C,D shows the partially optimized structures of BILP-3@3SO 2 and BILP-3@6SO 2 using the M06/6-311+G* level of theory.As expected, SO 2 molecules in BILP-3@3SO 2 preferably lie closer to the N-imine sites of the imidazole rings with an average sulfur−nitrogen distance of 2.58 Å, the same as that observed in fully optimized BILP-4@2SO 2 .When six SO 2 molecules interact in BILP-3@6SO 2 , three of them preferably bind to the N-imine sites, as expected, with an average sulfur−nitrogen distance of 2.42 Å.The remaining three SO 2 molecules lie close to the H atom of aryl-H that is located at the N-imine sites rather than the N−H site of imidazole moieties (Figure 4D).This result was surprising because it was expected that the later three SO 2 molecules were likely to lie toward the N−H sites.This spatial localization of SO 2 molecules can be explained by combining the effect of a higher electronegative charge around the benzene ring, which creates aryl-H partially positive (aryl-H(δ + )•••O(δ − )�S�O) and the cooperativity between SO 2 (I) and SO 2 (II) molecules (intermolecular O�S�O(δ − )•••S(δ + )-O 2 ).This combined attractive force appears to overcome the attraction between SO 2 (II) and N−H of the imidazole ring (O�S�O(δ − )•••H(δ + )N).However, the attractive force between SO 2 (II) and N−H of the imidazole ring in BILP-4@4SO 2 works together with the other two combined forces because all are closely spaced on the same side.This is supported by the fact that the binding energy (BE) of BILP-4@4SO 2 (42.31 kJ mol −1 ) is slightly higher than that of BILP-3@6SO 2 (40.31 kJ mol −1 , Table 1).The calculated binding energy for BILP-3@6SO 2 (40.31 kJ mol −1 ) using M06 supports the experimental heat of adsorption.Experimental Q st for BILP-3 at zero coverage, namely, 32.0 kJ mol −1 , increases with SO 2 uptake to a maximum of 38.3 and then gradually drops to 25.0 kJ mol −1 at 1.0 bar (Figure 2C).The increase in the Q st value with the increasing SO 2 uptake appears to be due to the cooperative nature of the adsorbed SO 2 molecules.Similar guest−guest interaction between two adjacent SO 2 molecules has been reported for MOF MFM-520 and for fluorinated anion-pillared metal−organic frameworks (APMOFs). 42,75Higher binding energy in LDA compared to M06 level of theory as summarized in Table 1 accounts for the overestimation of the binding in the LDA method. 32Our experimental and calculated binding energies from the M06 level of calculations are in good agreement with the binding energy reported for imidazole (39.1 kJ mol −1 ) by Shannon et al. 76 A similar result (40.52 kJ mol −1 ) was reported for SO 2 − tetrazine interaction based on the Monte Carlo simulation. 77ther reports on SO 2 capture by azole-based ionic liquids, derived from tetrazole and imidazole moieties, demonstrated that the electronegative nitrogen atoms of the anions can modulate bonding strength and lead to high absorption of SO 2 . 78The reported calculated enthalpies for SO 2 −tetrazole complexes were 89.3, 59.9, 39.7, and 34.4 kJ mol −1 for [Tetz] @SO 2 , [Tetz]@2SO 2 , [Tetz]@3SO 2 , and [Tetz]@4SO 2 , respectively.Similar results were also reported for the SO 2 − imidazole interactions [Im]@SO 2 (124.6 kJ mol −1 ), [Im]@ 2SO 2 (75.7 kJ mol −1 ), [Im]@3SO 2 (36.9 kJ mol −1 ), and [Im] @4SO 2 (30.3 kJ mol −1 ).The gradual decrease in enthalpies in these series was explained by the change of interactions from chemical to physical types. 78The closest calculated binding energy among the cage molecules is for CC3 (49.7 kJ mol −1 ), which contains imine nitrogen, while the other case molecules have significantly higher binding energies (86.4 kJ mol −1 for RCC3, which contains a secondary amine; 68.6 kJ mol −1 for FT-RCC3, which contains tertiary amine). 16Higher binding energy for amine (−NH 2 ) functionalized MOF NH 2 -MIL-53(AL) was reported to be 67.3 kJ mol −1 , which is dominated by the strong interactions of SO 2 with two NH 2 groups from two different directions (N−H 79 Based on the above discussions, it is expected that BILPs, which have neutral imidazole moieties in their backbones, exhibit only physisorption interactions with SO 2 .The calculated binding energy using the M06 method for BILPs@nSO 2 not only supports the physisorption assumptions but also supports the superiority of M06 over LDA at least when the adsorption is physisorption type.
Selective Gas Adsorption.Because of its high toxicity, trace levels of SO 2 must be captured before releasing flue gas or other SO 2 -containing gas mixtures into the atmosphere.Selective adsorption is essential for the effective capture and separation of trace levels of SO 2 from gas mixtures.6][27][28]80 This encouraged us to evaluate the efficiency of BILPs in the selective capture of SO 2 from common gas mixtures that contain trace levels of SO 2 . Theadsorption isotherms were fitted to dual site for SO 2 , CO 2, and CH 4 and single site for N 2 using Langmuir Freundlich models

Langmuir
(Figure 5).The fitted isotherms were used to calculate the selectivity based on the initial slope calculations (Figure S12).Significantly high SO 2 uptake capacities of BILP-3 and BILP-4 were observed compared to our reported CO 2 , CH 4 , and N 2 uptakes (Figures 5 and S12).31).Although the adsorption of SO 2 on BILPs has been attributed to the physisorption type as discussed in earlier sections, the relative adsorption affinity onto BILPs, which have basic nitrogen sites, is expected to be higher for SO 2 than that for CO 2 because of the higher acidic nature of SO 2 .This is further supported by their Q st values.The Q st values for CO 2 and CH 4 for BILP-3 at zero coverage are 28.6 and 16.6 kJ mol −1 , respectively, and they decrease steadily with increasing the gas uptakes. 26A relatively higher Q st value of 32.0 kJ mol −1 and the cooperative nature of adsorbed SO 2 provide higher adsorption selectivity for SO 2 over that of CO 2 at low concentrations.Furthermore, SO 2 is a polar molecule while CO 2 is nonpolar, and this polarity difference plays the key role in the preferential adsorption of SO 2 through dipole− dipole interaction.Nonpolar N 2 as the inert gas has the lowest affinity in the adsorption scale.These findings indicate the potential of BILPs in selective SO 2 removal from gas mixtures, where its concentration is very low.

■ CONCLUSIONS
We demonstrated that BILPs show a very high uptake of acidic SO 2 gas.These porous polymeric frameworks are regenerated easily and can be reused without losing their initial gas uptake capacity.The moderate heat of adsorption confirms relatively strong site-selective physisorption as predicted by DFT calculations.Initial slope calculations predict that the BILPs would be effective in the selective capture of SO 2 at low concentrations from flue gas components.

■ ASSOCIATED CONTENT
* sı Supporting Information