Tetracarboxylate Linker-Based Flexible CuII Frameworks: Efficient Separation of CO2 from CO2/N2 and C2H2 from C2H2/C2H4 Mixtures.

We report the synthesis, structure, and adsorption properties of two new metal-organic frameworks (MOFs) {[Cu2(bpp)3(L1)]·(bpp)·(4H2O)} (1) and {[Cu2(bipy)2(L2)(H2O)2]·(bipy)·(5H2O)} (2) obtained from two different flexible tetracarboxylate linkers (L1 and L2) of variable lengths and flexibility. While 1 comprising CuII, L1, and 1,3-bis(4-pyridyl) propane (bpp) is a 2D MOF with a cage-type structure, 2 consisting of CuII, L2, and 4,4'-bipyridine (bipy) has a 3D twofold interpenetrated structure. Both frameworks manifest permanent porosity, as realized from CO2 adsorption at 195 K. 2 shows excellent CO2/N2 and C2H2/C2H4 adsorption selectivity at 298 K. This has been established by using 2 as a separating medium in a breakthrough column for separating mixtures of CO2/N2 (15:85, v/v) and C2H2/C2H4 (1:99, v/v). The selectivity of 2 toward CO2 over N2 and C2H2 over C2H4 is governed by favorable thermodynamic interactions owing to its structural flexibility, unsaturated metal sites, and polar carboxylate groups. Thus, 2 proves to be an extremely efficient material for specific gas separation.

Introduction

Adsorptive separation[1] of a particular gas from a mixture using porous materials is a technique of supreme importance in recent times because of its relevance in industrial as well as environmental applications.[2,3] Different gas molecules are essential to sustain life[4] as well as to cater to our various energy requirements.[5] Thus, the effective sieving of the different components of a gas mixture is of paramount importance.[6] Hydrocarbons, particularly ethylene obtained from natural and shale gases, are extremely important in the polymer industry and to produce several other chemicals. During steam cracking for the synthesis of ethylene, acetylene (about 1%) is usually produced as a side product. This trace amount of acetylene is a contaminant which, if not removed, acts as a catalyst poison in ethylene polymerization and degrades the quality of the resultant polyethylene.[7] Thus, the development of porous materials which can facilitate selective adsorption of acetylene over ethylene is of paramount importance as it can show a cost-effective facile route for the challenging task of gas separation,[7−9] thereby helping the polymer industry immensely.

Metal–organic frameworks (MOFs),[10] well-known for their multifaceted applications in gas storage,[2,52] catalysis,[11] luminescence,[12] and sensing properties,[13,14] are extremely relevant in this regard. MOFs are much more efficient, both in terms of energy and cost, than traditional methods for gas separation, for example, cryogenic distillation or partial hydrogenation. Selectivity in MOFs can be achieved by tuning its pore size by modulating the functionalities of the organic linker along with implementation of suitable metal ions with varied geometry.[15−17] By controlling the window dimension or pore size of MOFs, kinetic-based separation can be achieved.[2,18,19] Also, separation can be done on the basis of favorable thermodynamic interaction of the specific adsorbate molecules with the pore surface.[8,20,21] Thus, it is extremely important to adopt a proper targeted ligand design strategy to synthesize such MOFs with high efficiency. Multidentate flexible linkers, for example, di, tri, and tetra carboxylates,[24−27,51] are the most widely used linkers in MOF chemistry.[22,23] Such linkers coordinate with metal ions to form a wide range of multinuclear nodes with predictable geometries such as binuclear paddle-wheel units, tri, or tetranuclear units.[25] In this regard, aromatic tetracarboxylate linkers are particularly utilitarian because of the more number of coordinating sites as well as the rich π-electron density.[25,27] MOFs constructed from such π-electron-rich tetracarboxylate ligands have pore surfaces studded with a large number of oxygen atoms providing better interactions with gas molecules such as CO2[28,29] or small hydrocarbon molecules such as ethylene and acetylene.[30,31] Furthermore, the flexibility of such ligands contributes to the tunable porosity of MOFs by inducing guest-responsive structural dynamicity[23,31,32,50] with the ability to respond to external stimuli,[33] for example, temperature, pressure, light, electric fields, and chemical inclusion.[11,32] Such dynamic frameworks[34] also manifest stepwise[35] or gated adsorption[36,37] at a certain pressure of an adsorbate molecule[38,39] which results in guest selectivity[31,40] and gas separation.[40−42]

In this work, we have used two different flexible tetracarboxylate π-electron-rich linkers [5,5′-(1,4-(2-butene)bis(oxy)diisophthalic acid (L1) and 5,5′-(1,4-phenylenebis(methylene))bis(oxy)diisophthalic acid (L2) (Scheme ) of different lengths and flexibility. Self-assembly of L1 and 1,3-bis(4-pyridyl) propane (bpp) with CuII at room temperature results in a hydrophobic 2D MOF {[Cu2(bpp)3(L1)]·(bpp)·(4H2O)} (1) (CCDC no. 1590169) with a cage[43]-type structure. The permanent porosity of the framework has been established by performing CO2 adsorption at 195 K by different vapor adsorption studies. On the other hand, self-assembly of L2 and bipyridine (bipy) with CuII resulted in a twofold interpenetrated 3D MOF {[Cu2(bipy)2(L2)(H2O)2]·(bipy)·(5H2O)} (2) (CCDC no. 1590168). The desolvated framework (2′) exhibits an excellent selectivity toward C2H2 over C2H4 at 298 K, the practical applicability of which has been proven using a breakthrough column experiment. Additionally, it is also a selective adsorbent of CO2 from its mixture with N2, making it a potential material for the separation for industrial flue gas. In both cases, the basis of separation is the favorable thermodynamic interaction of the framework with respective adsorbate molecules.

Scheme 1

pH-Dependent Synthesis of Compounds 1 and 2 by the Self-Assembly of CuII and Respective Linkers at Room Temperature

Results and Discussion

Description of the Crystal Structure of Compounds 1 and 2

Single-crystal X-ray diffraction (SCXRD) data reveal that compound 1 crystallizes in a monoclinic P21/c space group. The asymmetric unit contains two crystallographically independent penta-coordinated CuII centers that have a (4 + 1) square pyramidal geometry. The degree of deviation from ideal square pyramidal geometry has been calculated in terms of Addison parameter (τ),[44] the value of which is merely 0.019 for Cu1, suggesting a square pyramidal geometry rather than a trigonal bipyramid. The basal plane of Cu1 is occupied by N3 and N6 from two different bpp linkers along with two carboxylate oxygen atoms O1 and O4 from two different L1 linkers. The apical position is occupied by the N1 atom from another bpp ligand (Figure a). The Cu1–O1 bond length at the basal plane is 1.93(8) Å, whereas that for Cu1–O4 is 1.97(8) Å. The bond lengths for Cu1–N3 and Cu1–N6 are 2.03(10) and 2.04(14) Å, respectively. However, the Cu1–N1 bond at the apical position is much longer with a bond length of 2.32(10) Å, implying a (4 + 1) geometry. Similarly, the pentacoordination of Cu2 is fulfilled by two carboxylate oxygen atoms (O5 and O7) from L1 and two nitrogen atoms (N4 and N5) from two different bpp linkers at the basal plane and N2 at the apical position (Figure a). The τ value for Cu2 is 0.13, suggesting high distortion from ideal square pyramidal geometry. The bond lengths of Cu2–O5 and Cu2–O7 are 1.97(10) and 1.94(9) Å, respectively, whereas those of Cu2–N4 and Cu2–N5 are 2.03(12) and 2.04(12) Å, respectively. The apical Cu2–N2 is much longer, with a bond length of 2.36(13) Å. Cu1 and Cu2 are connected together by a bpp linker to form a wavy chain (Figure b). Two such parallel chains are connected by an L1 linker, where the two carboxylates at one terminal of L1 connect two Cu1 atoms and the two carboxylates at the other terminal connect two Cu2 atoms (Figure d). Adjacent to this, the Cu1 center is further diagonally connected to a Cu2 atom in the other chain by a bpp linker. Thus, a 2D network with a cagelike structure is formed containing a hydrophobic pocket along the b direction (Figure d), which is lined by the π-electron-rich aromatic L1 on one side and a bpp linker on the other. This hydrophobic pocket houses a guest bpp molecule which forms a π–π interaction with the pyridyl ring of the bpp ligand on the wall of the pocket (cg–cg distance 4.6 Å). The central C3 alkyl moiety in the case of both L1 as well as bpp imparts flexibility to the pockets, thereby ensuring uptake of adsorbate molecules in spite of the small pore size. All the guest water molecules present in the framework, except for O5w, are hydrogen-bonded with the framework. The hydrogen-bonding parameters of 1 have been summarized in Table S3. The solvent accessible void space in the framework is 2943.3 Å3 (39.6% of the total cell volume), which has been calculated using PLATON after removing the guest bpp and the water molecules.

Figure 1

(a) Coordination sphere of Cu1 and Cu2 in compound 1. (b) 1D wavy chain of 1 formed by Cu centers linked with bpp molecules. (c) 2D network in 1. (d) Hydrophobic pockets in the framework.

Compound 2 crystallizes in the C2/c monoclinic space group. Each asymmetric unit comprises two crystallographically independent pentacoordinated CuII centers, each of which assumes a distorted square pyramidal geometry, the degree of distortion being as low as 0.007 for Cu1 and 0.037 for Cu2. The pentacoordination of Cu1 is fulfilled by two nitrogen atoms (N1 and N2) from two different bipy molecules, two oxygen atoms (O1 and O3) from tetracarboxylate linker L2, and one coordinated water molecule (O7w). Similarly, the pentacoordination of Cu2 is satisfied by two nitrogen atoms (N3 and N6) from two different bipy linkers, two oxygen atoms (O5 and O12) from L2, and one oxygen atom (O5w) from a coordinated water molecule. The bipy linker (containing N6 and N7 atoms), connecting two adjacent Cu2 centers, is disordered over two positions with 50–50 equal occupancy (Figure S5a). Cu1 centers connected by bipy linkers form a 1D chain along the a direction, and a similar 1D chain is formed by the bridging of Cu2 centers by bipy along the crystallographic a direction (Figure a). The two carboxylates on one terminal of an L2 linker (marked in violet in Figure b) connect two of these parallel chains on one plain (marked in dark green in Figure b). The two carboxylates on the other terminal of the L2 connect two other such parallel 1D chains on a different plain (marked in blue in Figure b) to form the 3D framework (Figure c). Two 3D frameworks interpenetrate each other, resulting in a twofold interpenetrated 3D framework (Figure d). Along the crystallographic b direction, the single nets of the 3D framework contain a 1D channel with a pore dimension of 4.45 × 5.46 Å2, which upon twofold interpenetration (Figures c and S3b) reduces to a smaller channel having a pore dimension of 2.93 × 5.14 Å2, and the channel is occupied by guest bipy and guest water molecules. Detailed analysis with TOPOS[48,49] reveals that compound 2 is the 5-nodal (2-c)(4-c)2(4-c)2 periodic net connected by two 4-connected (4-c) Cu centers, by two 4-connected (4-c) L2 ligands, and one 2-connected (2-c) bipy ligand. Further analysis shows that compound 2 adopts an unprecedented network topology with the Schlafli symbol {62.72.8.9}2{63.7.82}{64.8.9}{64.82}2{7}. All the water molecules, except O4w, are hydrogen-bonded to the framework, and the hydrogen bonding parameters have been summarized in Table S6. The pyridyl ring (containing the N4 atom) of guest bipy molecules shares a π–π interaction (cg–cg distance 3.76 Å) with the pyridyl ring (containing the N1 atom) of bipy linkers along the pore wall. As we have discussed earlier, both Cu1 and Cu2 are coordinated to water molecules, and thus, unsaturated metal sites (UMS) could be generated after desolvation. The solvent accessible void is 5826.1 Å3 (47.5% of the total cell volume) and has been calculated using PLATON after removing the coordinated and as well as guest water and guest bipy molecules.

Figure 2

(a) 2D sheet viewed along the crystallographic c direction in 2. (b) 3D network of 2. (c) 3D two- fold interpenetrated framework of 2. (d) Pore structure in 2. (e) Pendent oxygen atoms and unsaturated CuII sites on the pore surface of 2.

Framework Stability and PXRD

The thermogravimetric analysis (TGA) of compound 1 shows an initial loss of ∼4%, corresponding to three guest water molecules (Figure S6a). Subsequently, there is a continuous weight loss beyond ∼150 °C which may be attributed to the loss of the remaining two guest water molecules and the guest bpp molecule, which can be explained on the basis of SCXRD analysis. Two of the guest water molecules are released at higher temperature. Beyond 200 °C, the framework degrades continuously.

The powder X-ray diffraction (PXRD) pattern of the as-synthesized compound 1 matches perfectly with the simulated pattern, suggesting the phase purity (Figure a). The framework on being desolvated undergoes structural changes, as evident from the PXRD pattern of the framework desolvated at 135 °C under vacuum (10–1 Pa). On desolvation, there is rearrangement in the Bragg’s reflections, along with peak broadening and appearance of new peaks, suggesting a structural transformation.

Figure 3

(a) PXRD pattern of 1 in different states. (b) PXRD pattern of 2 with the increasing temperature. (c) CO2 and N2 adsorption isotherms of compound 1. (d) Solvent adsorption isotherms of compound 1. (e) CO2 adsorption isotherm of 2′ at 195 and 273 K. (f) Solvent adsorption isotherms of 2′.

The TGA of 2 shows an initial loss of ∼9% up to 100 °C owing to the loss of all 5 guest water molecules and one of the coordinated water molecules (Figure S6a). The subsequent ∼13% loss starting at 140 °C can be attributed to the loss of the second coordinated water molecule and guest bipyridine molecules. The desolvated framework degrades almost continuously beyond this.

The PXRD pattern of the as-synthesized compound 2 is in accordance with the simulated pattern. The structure of compound 2 shows changes with the increasing temperature, as evident from the desolvated PXRD pattern (Figure b). This can be attributed to the removal of the coordinated water molecules and the guest bipy molecules, which interact with the framework via π–π stacking. The removal of the bipy molecule gives rise to another structural phase. In this framework, along with guest water molecules, there are coordinated water molecules, and their removal results in the generation of UMS. Such a change in the coordination sphere of the CuII center ensues structural changes, as reflected from the PXRD pattern.

Gas and Solvent Vapor Adsorption

Although compound 1 does not have a definite pore in any direction, the presence of a pocket lined by electron rich aromatic benzene and pyridine rings along the crystallographic b direction (Figure d) prompted us to check the adsorption behaviour of the desolvated compound (1′). At 195 K, the adsorption isotherm of 1′ shows an uptake of 35 mL g–1 CO2, with a steep uptake in the low-pressure region. This can be attributed to the interaction of the CO2 molecules with the free pendant carboxylate oxygen atoms flanking the pore surface and the π-electron density on the pore surface from the aromatic rings of the ligand (Figure c). The water adsorption curve (Figure S7a) confirms the hydrophobic nature of the framework exhibiting poor uptake at lower pressures, which is also reiterated in the MeOH and EtOH adsorption isotherms (Figure d). The net adsorption amount decreases from water (2.7 Å) to methanol (3.6 Å) to ethanol (4.5 Å) as the kinetic diameter of the solvent molecules increase in that order. The lower polarity of MeOH ensures a steeper uptake at low pressures, as compared to water, but the net uptake of water is much higher owing to its smaller size. Although the hydrophobic MOF is expected to show a good adsorption of nonpolar solvents such as benzene, the small pore dimensions of 1 restrict the adsorption of benzene (kinetic diameter of benzene being 5.9 Å) (Figure S7a).

To check the porous properties of compound 2, the N2 adsorption isotherm of the desolvated framework (2′) was measured at 77 K. 2′ does not adsorb N2, which is attributed to the high-diffusion barrier of the inert N2 gas (kinetic diameter: 3.65 Å) in the narrow channel of 2′ (2.93 × 5.14 Å2 along the b direction) (Figure d). Moreover, the accessible void space in this framework is very small because of the twofold interpenetration. However, framework 2′ manifests an appreciable amount of CO2 uptake at 195, 273, as well as at 298 K. At 195 K, 2′ records the adsorption of 127 mL g–1 CO2, with a steep uptake in the low-pressure region, indicating its microporous nature (Figure e). Steps are observed at various stages of adsorption, which manifest that the framework is flexible in nature, and there are different adsorption sites for CO2. Desorption does not follow the same path as adsorption initially, and we observe a hysteresis till P/P0 = 0.55, below which both the curves converge. The adsorption amount of CO2 is 85 mL g–1 at 273 K and 65 mL g–1 at 298 K. 2′ shows type-I adsorption at 298 K where the adsorption and desorption curves do not coincide, exhibiting a narrow hysteretic profile (Figure e). Besides the pendant oxygen atoms on the pore surface, the presence of UMS, that is, pentacoordinated CuII centers, along the pore further contributes to the enhanced adsorption (Figure e). Also, the kinetic diameter of CO2 is smaller (3.64 Å), and its high quadruple moment interacts with the π-electron cloud of the aromatic rings on the pore surface. The polar hydrophilic nature of the pore surface of 2′ is manifested by the solvent adsorption profiles (Figures f and S5b). The adsorption isotherms of CO2 at 273 and 298 K were fitted in the virial equation, from which the isosteric heat of adsorption (Qst,f) for 2′ was calculated to be ∼18.9 kJ mol–1 at the zero coverage region (Figure S10b).[8]

Intrigued by this, we proceeded to predict the selectivity of 2′ toward CO2 over N2 at 298 K using the Clausius–Clapeyron equation. At 298 K for 2′, the predicted ideal adsorbed solution theory (IAST)[45] selectivities of CO2 over N2 for a bimolar (0.15:0.85) mixture of CO2–N2 is 13. Also here, the selectivity increases with the increasing pressure. At pressure 101 kPa, the CO2–N2 selectivity is 94 (Figure S9a).

Such a good theoretical value prompted us to check the actual separation ability of 2 for CO2 gas from a binary CO2/N2 (0.15:0.85) mixture at room temperature (293 K). Tightly packed columns of 2′ (400 mg, column length 4 cm) were prepared and activated with a He gas (flow rate 1 mL/min). The breakthrough curve of the binary mixture of CO2/N2 (0.15:0.85) for 2′ is represented in Figure c, where it exhibits a CO2 retention time of 11 min. This high retention time indicates that 2′ holds immense potential for practical use as a medium for separating CO2 from other gases.

Figure 4

(a) CO2 adsorption isotherm at 298 K and N2 at 77 K for 2′. (b) Adsorption isotherms for C2H2 and C2H4 at 293 K for 2′. Breakthrough curves for a binary mixture of (c) N2/CO2 (0.85:0.15) and (d) C2H2/C2H4 (0.1:0.99) for 2′.

Inspired by such efficiency of 2′ in separating CO2 and N2, we were further interested in exploiting the flexible structure and UMS for C2 [acetylene (C2H2) and ethylene (C2H4)] hydrocarbon storage and separation. Thus, 2′ was checked for C2H2 and C2H4 adsorption, and it was found that 2′ can adsorb 38 mL g–1 C2H2 at 273 K, whereas the adsorption amount for C2H4 is 27 mL g–1 at 273 K (Figure S8). 2′ shows a steeper uptake of C2H2 at lower pressure, as compared to C2H4. At P/P0 = 0.01, the adsorbed amount of C2H2 per gram of 2′ is 11.5 mL when the adsorbate is C2H2, whereas that for C2H4 is merely 5 mL. The desorption curve for C2H2 takes a different route than the adsorption isotherm, thereby forming a hysteresis, indicating a strong interaction between the framework and C2H2. At 293 K, the adsorption amounts are 33 mL g–1 for C2H2 (Figure b) and 16 mL g–1 for C2H4. The isosteric heat of adsorption calculated for C2H2 is 32 kJ mol–1 at the zero coverage region (Figure S8d,e). Thus, because of a higher Qst value, C2H2 attains more accessibility to the pore surface and interacts with the UMS, unlike C2H4. At 293 K, the predicted selectivity of C2H2 over C2H4 for a (0.01:0.99) mixture of the two gases is calculated in a similar manner as for CO2–N2, using IAST,[45] which is 3.7 initially, and the selectivity decreases gradually with increasing pressure (Figure S9b). Thus, the easy accessibility to UMS coupled with good IAST selectivity, high room-temperature uptake, and relatively low heat of adsorption certifies that 2′ is an ideal material for C2H2–C2H4 separation.

To establish the practical applicability of 2 in separating a trace amount of C2H2 from C2H4, we performed an actual breakthrough experiment where we fed a column of 2′ with a (1:99, v/v) mixture of C2H2 and C2H4 at 298 K. Interestingly, C2H4 appears within a few seconds of starting the gas flow. The breakthrough time for C2H2 is 27 min (Figure d). This is a rather unique feat achieved by any MOF at room temperature and aptly upholds the efficiency of 2 in real-time applications.

Conclusions

In conclusion, two different flexible tetracarboxylate ligands L1 and L2 have been successfully implemented in two different frameworks, 1 and 2, along with other linkers. The variable flexibility and the aromatic π-electron-rich nature of the ligands impart very interesting properties within the framework. The presence of a hydrophobic pocket, lined by aromatic rings from the ligand in compound 1, enables it to adsorb a reasonable amount of CO2 at 195 K. On the other hand, compound 2, with a 3D framework with twofold interpenetration, is a promising adsorbent material owing to the presence of the polar carboxylate groups, UMS, and flexible structure. It shows a good IAST selectivity toward C2H2 over C2H4, which is culminated in practical applications using a breakthrough column. 2′ can retain C2H2 for a very long time within this breakthrough column, releasing pure C2H4. Moreover, 2′ also manifests excellent CO2–N2 separation. The selectivity of CO2 over N2 and C2H2 over C2H4 is attributed to more favorable thermodynamic interactions. Thus, frameworks constructed from such tetracarboxylate ligands can be used as effective materials for specific gas sequestration and separation by tuning flexibility and functionalities.

Experimental Section

Materials

All the reagents and solvents were used as obtained from commercial supplies without any further purification. Dimethyl-5-hydroxy isophthalate, α,α′-dibromoparaxylene, dibenzo-18-crown-6, 1,4-dibromobutene, and K2CO3 were obtained from Aldrich Co. Ltd.

Synthesis of {[Cu2(bpp)3(L1)]·(bpp)·(4H2O)} (1)

CuCl2·2H2O (0.2 mmol, 0.034 g) and solid L1 (0.1 mmol, 0.047 g) were mixed with 10 mL of water and stirred for 10 min. Aqueous ammonia (20%) was added dropwise to this mixture till the solution becomes clear. To this, 10 mL methanolic solution of 1,3-bis(4-pyridyl)propane (0.5 mmol, 0.099 g) was added, and the resulting solution was stirred for 1 h. The solution was kept undisturbed at room temperature. After 3–4 days, dark blue needle-shaped crystals appear. Yield: 60% relative to CuII. Anal. Calcd for C12H17Cu2N17: C, 61.051; H, 5.123; and N, 7.910%. Found: C, 60.99; H, 5.43; and N, 7.72%. FT-IR (4000–400 cm–1) 1568 (s), 1618 (s), 1358 (s), and 778 (s) cm–1.

Synthesis of {[Cu2(bipy)2(L2)(H2O)2]·(bipy)·(5H2O)} (2)

CuCl2·2H2O (0.2 mmol, 0.034 g) and solid L1 (0.1 mmol, 0.047 g) were mixed with 20 mL of water and stirred for 10 min. Aqueous ammonia (20%) was added dropwise to this reaction mixture until the solid residues dissolve, and the solution becomes clear. Subsequently, a 10 mL methanolic solution of 4,4′-bipyridine (0.5 mmol, 78 mg) was added to the previous reaction mixture, and the solution was stirred overnight. The solution was kept undisturbed at room temperature. Light blue crystals are formed after 3–4 days. Yield: 72% relative to CuII. Anal. Calcd. for C54H46Cu2N6O: C, 55.054; H, 3.935; and N, 7.133%. Found: C, 55.26; H, 3.45; and N, 7.27%. FT-IR (4000–400 cm–1) 782 (s), 1265 (s), 1354 (s), 1570 (s), 1616 (s), and 3900 (w) cm–1.

For characterization and adsorption studies, phase pure single crystals were used.

Physical Measurements

The elemental analyses were carried out using a Thermo Fischer FLASH 2000 elemental analyzer. IR spectra were recorded with a Bruker IFS 66v/S spectrophotometer using KBr pellets in the region of 4000–400 cm–1. TGA was carried out (Mettler Toledo) in a nitrogen atmosphere (flow rate = 50 mL/min) in the temperature range of 30–500 °C (heating rate = 3 °C min–1). PXRD patterns of the compounds in different states were recorded by using Cu Kα radiation (Bruker D8 DISCOVER; 40 kV, 30 mA).

X-ray Crystallography

X-ray single-crystal structural data for 1 and 2 were collected using a Bruker SMART-CCD diffractometer equipped with a normal focus and a 2.4 kW sealed tube X-ray source with graphite monochromated Mo Kα radiation (λ = 0.71073 Å) operating at 50 kV and 30 mA. The program SAINT[14] was used for integration of diffraction profiles, and absorption correction was made using the SADABS[15] program. All of the structures were solved using SHELXT[47] and refined by a full matrix least square method using SHELXL-2017.[46] All of the hydrogen atoms were geometrically fixed and placed in ideal positions. The potential solvent accessible area or void space was calculated using the PLATON multipurpose crystallographic software.[18] All crystallographic and structure refinement data for 1 and 2 are summarized in Table . Selected bond lengths and angles for 1 and 2 are given in Tables S1–S6, respectively.

Table 1

Crystal Data and Structure Refinement Parameters for Compounds 1 and 2

parameter	1	2
empirical formula	C72H72Cu2N8O15	C54H46Cu2N6O17
formula weight	1398.4	1177.9
crystal system	monoclinic	monoclinic
space group	P21/c (no. 14)	C2/c (no. 15)
a, Å	23.6727(17)	44.8511(10)
b, Å	18.9130(13)	19.0460(4)
c, Å	17.6128(13)	14.7115(3)
β, deg	109.696(3)	102.380(1)
V, Å3	7424.6	12 274.8
Z	4	8
T, K	293	293
μ, mm–1	0.638	0.766
Dcalcd, g/cm3	1.249	1.331
F(000)	2904	5072
reflections [I > 2σ(I)]	3751	6004
unique reflections	8042	9461
measured reflections	94 274	81 471
Rint	0.171	0.085
GOF on F2	0.96	1.03
R1[I > 2σ(I)]a	0.1040	0.068
Rw[I > 2σ(I)]b	0.2956	0.2170

R = ∑||F0| – |Fc||/∑|F0|

Rw = [∑{w(F02 – Fc2)2}/∑{w(F02)2}]1/2.

Adsorption Study

The adsorption isotherms of CO2 (195, 273 and 298 K) and N2 (77 K, 298 K), and other gases (C2H2 and C2H4 at 273 and 298 K) using the desolvated sample of 1 (1′) and 2 (2′) were measured by using a Quantachrome Autosorb IQ2 analyzer. In the sample tube adsorbent, samples 1 and 2 (∼100–150 mg) were placed which had been prepared at 353 and 433 K, respectively, under a 1 × 10–1 Pa vacuum for about 6 h prior to the measurement of the isotherms. Helium gas (99.999% purity) at a certain pressure was introduced in the gas chamber and allowed to diffuse into the sample chamber by opening the valve. The amount of gas adsorbed was calculated readily from the pressure difference (Pcal – Pe), where Pcal is the calculated pressure with no gas adsorption and Pe is the observed equilibrium pressure. All operations were computer-controlled and automatic.

The adsorption of different solvents such as MeOH at 293 K and H2O and EtOH at 298 K was measured in the desolvated sample of 1 and 2 in the vapor state by using a BELSORP Aqua-3 analyzer. The samples of about ∼100–150 mg were activated under similar conditions as mentioned earlier. The different solvent molecules used to generate the vapor were degassed fully by repeated evacuation. The dead volume was measured with helium gas. The adsorbate was placed into the sample tube, and then the change of pressure was monitored and the degree of adsorption was determined by the decrease in pressure at the equilibrium state. All operations were computer-controlled and automatic.

Breakthrough Measurements

Breakthrough experiments were performed using a column (packed with about 400 mg of powdered 2′), 4 cm long in length and 0.3 cm in diameter. The sample was first activated at 160 °C for 12 h prior to loading in the column. After loading, the sample was again activated with the He flow for 30 min. CO2/N2 (v/v 15:85 mixture) and C2H2/C2H4 (v/v 1:99 mixture) streams were passed through the column with He. The flow was continuous, and the flow rate was regulated by a Mass Flow Controller (MFC). The partial pressure of the gas mixture was kept at p = 5 kPa. The gas stream at the outlet of the column was analyzed on-line with gas chromatography. The space velocity was kept at 4.71 min–1 (Figure S11).