Methane Storage in Paddlewheel-Based Porous Coordination Cages

Abstract

Although gas adsorption properties of extended three-dimensional metal−organic materials have been widely studied, they remain relatively unexplored in porous molecular systems. This is particularly the case for porous coordination cages for which surface areas are typically not reported. Herein, we report the synthesis, characterization, activation, and gas adsorption properties of a family of carbazole-based cages. The chromium analog displays a coordination cage record BET (Brunauer−Emmett−Teller) surface area of 1235 m²/g. With precise synthesis and activation procedures, two previously reported cages similarly display high surface areas. The materials exhibit high methane adsorption capacities at 65 bar with the chromium (II) cage displaying CH₄ capacities of 194 cm³/g and 148 cm³/cm³. This high uptake is a result of optimal pore design, which was confirmed via powder neutron diffraction experiments.

As a result of continually increasing global reserves,¹ natural gas has been touted as a cleaner alternative to gasoline and coal as a fuel in the transportation and power generation sectors, respectively.^2,3 Because natural gas primarily consists of methane, which is the upper limit of H:C ratio, it is the most gravimetrically energy dense hydrocarbon fuel. Consequently, its combustion releases considerably less pollutants on a per energy basis.⁴ Indeed, U.S. CO₂ emissions from power generation are near a 30 year low resulting from a transition to natural gas.⁵ In the transportation sector, however, the low volumetric energy density of methane at STP has limited its utility.⁶ Proposed solutions to this have included liquefaction or compression. Both of these have generally been viewed as not viable in passenger vehicles.⁷ Adsorbed natural gas systems have shown promise in this regard.⁸ The challenge here is the development of an adsorbent with appropriate methane capacity. Zeo-lites,^9,10 activated carbons,¹¹ metal−organic frameworks (MOFs),^12–18 and a variety of other porous materials have been investigated for high pressure natural gas storage. MOFs currently display the most promising uptakes^19–22 with reported CH₄ adsorption capacities as high as 259 cm³/cm³.²³ Additionally, a number of interesting adsorption properties have been observed in these materials, including flexibility for thermal management and temperature-dependent negative gas adsorption.^22,24 However, advancements in this area remain short of Department of Energy targets.²⁵

Porous molecular adsorbents may show promise in this regard as they can be thought of as soluble adsorbents with designer pores. This could potentially endow them with favorable properties such as tailored syntheses, flexibility upon adsorption, cooperative binding, and increased ease of processability. These molecules, which can be either all-organic^26–28 or metal−organic,^29–31 are discrete zero-dimensional materials that display permanent porosity in the solid state. Although the underlying chemistry of porous organic cages and porous coordination cages has been established for decades, permanent porosity in these materials is a relatively recent phenomena when compared to zeolites and MOFs. As a result, they have remained almost completely unexplored for high-pressure gas storage applications. This is somewhat surprising given their conceptually analogous nature to porous extended solid materials. In fact, many metal−organic frameworks contain cages as their building units.^32,33 In an effort to expand the library of carboxylate-based porous cages, we have thoroughly investigated the synthesis of novel isophthalic acid and carbazole-dicarboxylic acid materials.³⁴ These latter molecules are particularly interesting for gas storage applications as the M₁₂(cdc)₁₂ (cdc²⁻ = carbazoledi-carboxylic acid) octahedral cage is present as one of three cages in DUT-49²⁰ and PCN-82,³⁵ two MOFs with high methane and hydrogen storage capacities, respectively. Given the high surface area of recently reported cuboctahedral M₂₄(R-bdc)₂₄ cages based on chromium(II),^30,36 we sought to expand the synthesis of M₁₂(cdc)₁₂ to this metal as a porous cage of this type is expected to show promising high-pressure adsorption properties.

Analogous to the synthesis of the Cu- and Mo-based M₁₂(cdc)₁₂ materials, the air-free reaction of anhydrous Cr₂(OAc)₄ with H₂cdc in a dimethylformamide (DMF)/methanol (MeOH) mixture affords Cr₁₂(cdc)₁₂·nDMF as large red/purple crystals. Single crystal X-ray diffraction confirms the material is isostructural to the previously reported Cu²⁺ and Mo²⁺ cages.^37,38 This structure is comprised of six bimetallic paddlewheel units in an octahedral arrangement coordinated to 12 cdc²⁻ ligands (Figure 1). The cages consist of 8 triangular windows with corner−corner distances of approximately 12 Å. The molecules pack in the solid state with cage center to center distances of approximately 22 Å. The structure of Cr₁₂(cdc)₁₂ contains solvent accessible voids in excess of 60%, which rivals the values displayed by many MOFs. Accordingly, thermogravimetric analysis (TGA) of a DMF-exchanged sample reveals a mass loss of 40% at 250 °C (Figure S4). Precise solvent exchange protocols must be implemented to achieve high surface areas for molecular materials as cages are potentially soluble in a variety of solvents. For Cr₁₂(cdc)₁₂, room temperature solvent exchanges with DMF and methanol were used. Powder X-ray diffraction confirms the as-synthesized material retains high crystallinity upon exchange with DMF and methanol (Figure 2). However, given the lack of three-dimensional connectivity and in contrast to most MOFs, there is significant structural rearrangement upon initial solvent exchange. Activation at 70 °C affords a BET surface area of 1235 m²/g. To the best of our knowledge, this is the highest surface area reported for a molecular metal−organic material.

Figure 1.

Structure of Cr₁₂(cdc)₁₂ as determined by single-crystal X-ray diffraction. Purple, gray, red, and blue represent chromium, carbon, oxygen, and nitrogen, respectively. Hydrogen atoms and solvent molecules have been omitted.

Figure 2.

Powder X-ray diffraction patterns of as-synthesized (black), DMF exchanged (red), MeOH exchanged (blue), and desolvated (green) Cr₁₂(cdc)₁₂ vs predicted (purple).

In order to achieve comparable surface areas for Cu₁₂(cdc)₁₂ and Mo₁₂(cdc)₁₂, we surveyed a range of synthesis, solvent exchange, and activation procedures. In contrast to metal− organic frameworks where identical surface areas for a material should be obtained regardless of synthesis conditions (assuming phase purity and full activation), a porous coordination cage can deposit in different crystalline phases depending on synthetic conditions. Although the individual cage structures remain in these various analogs, their three-dimensional packing, stability, and porosity can potentially vary. For Mo₁₂(cdc)₁₂, a DMPU-based synthesis was reported affording a solid that crystallizes in $R\overline{3}$.³⁸ Despite our best efforts, we were only able to obtain an activated material with a surface area of approximately 550 m²/g. Synthesis of the cage in DEF affords the same cage that instead crystallizes in P2₁/n. Methanol exchange of this material followed by activation at 50 °C affords a solid with a BET surface area of 1108 m²/g. In an analogous manner, Cu₁₂(cdc)₁₂ can be synthesized and crystallized from a variety of amide-based solvents. The highest surface area that was observed (657 m²/g) was for a material synthesized in DMF/MeOH, washed in MeOH, and activated at 50 °C.

To evaluate the methane storage potential of the three porous cages, high-pressure adsorption isotherms were collected from 0 to 65 bar at 298 K. At 35 bar the total gravimetric uptakes follow the surface area trends with capacities of 148, 135, and 81 cm³/g for Cr₁₂(cdc)₁₂, Mo₁₂(cdc)₁₂, and Cu₁₂(cdc)₁₂, respectively (Figure 3). These values are not only significantly higher than those for previously reported coordination cages,³⁹ they are consistent with the capacities displayed by metal−organic frameworks with similar surface areas.¹⁵ Methane uptake increases up to at least 65 bar with the Cr₁₂(cdc)₁₂ displaying the highest gravimetric capacity of 194 cm³/g while Mo₁₂(cdc)₁₂ has the highest volumetric capacity (150 cm³/cm³) (Table 1) (Figures S30, S31). To the best of our knowledge, both the gravimetric and volumetric capacities are the highest observed for a porous molecular assembly. For all three materials, moderate hysteresis is seen upon desorption, likely a result of the lack of three-dimensional connectivity between the cages.

Figure 3.

Total high-pressure CH₄ adsorption at 298 K in Cr₁₂(cdc)₁₂ (purple), Mo₁₂(cdc)₁₂ (orange), Cu₁₂(cdc)₁₂ (green), and PCN-81 (blue).

Table 1.

Methane Storage Properties of M₁₂(cdc)₁₂ Cages in Comparison to PCN-81, HKUST-1, and IRMOP-51^a

Material	BET Surface Area m2/g	35 bar Capacity cm3/g [cm3/cm3]	65 bar Capacity cm3/g [cm3/cm3]
IRMOP-5139	480	48.4 [25]	—
Cu12(cdc)12	657	81.3 [69.4]	111.8 [95.4]
Cr12(cdc)12	1235	147.5 [112.1]	193.9 [147.8]
Mo12(cdc)12	1108	135.0 [111.5]	181.1 [150.0]
PCN-81	4050	305.3 [146.8]	448.7 [215.8]
HKUST-115	1850	255.24 [224.9]	298.5 [263.0]

^aAll values are total uptake.

To directly compare the CH₄ storage capacities of M₁₂(cdc)₁₂ cages with MOFs containing similar methane binding environments, we targeted previously reported materials with M₁₂(cdc)₁₂ as their building units. DUT-49 has been thoroughly investigated for its high CH₄ uptake at high pressure, however the flexibility of the material complicates adsorption studies.²⁴ Two analogous materials, PCN-81 and PCN-82, are similarly based on M₁₂(cdc)₁₂ cages, although the former was shown to lack permanent porosity.³⁵ However, synthesis of PCN-81 via the reported route followed by thorough room temperature DMF and MeOH exchanges and activation at 100 °C afforded a highly crystalline material with a BET surface area of 4050 m²/g. Consistent with its high gravimetric surface area, PCN-81 displays an incredibly high CH₄ uptake of 305 cm³/g at 35 bar and 298 K. At 65 bar this value reaches 449 cm³/g. Given the low density of the material, its volumetric capacity at 35 bar is significantly lower at 147 cm³/cm³. As a result of the shallow nature of its adsorption isotherm (Figure 3), PCN-81 has a deliverable CH₄ capacity of 190 cm³/cm³ for a pressure swing of 65 to 5 bar, which is on par with the highest value reported for a porous material. Low temperature isotherms collected over a range of temperatures indicate methane adsorption enthalpies of −16 to −18.5 kJ/ mol for all four materials.

We turned to powder neutron diffraction to gain insight into the nature of CH₄ adsorption in these materials. Although the porous coordination cages retain crystallinity upon solvent removal, it is decreased during activation to the point of precluding the use of diffraction studies to interrogate binding sites. PCN-81, however, remains highly crystalline upon degassing. Refinement of data collected on an activated sample reveals it is significantly less distorted upon evacuation (Figure 4) compared to its solvated state.³⁵ At a CD₄:Cu loading of 1:1, three main adsorption sites are apparent. Although the framework is composed of three types of pores, the methane is exclusively adsorbed in the M₁₂(cdc)₁₂ cage. A primary adsorption site is the open copper(II) center with a Cu−C distance of 2.777(5) Å and an occupancy of 0.63(3). A site with similar occupancy (0.64(3)) is at the edge of a carbazole ligand with C_methane−C_ligand distances of 3.27 and 3.69 Å. This methane molecule is only present in six of the eight triangular windows of the Cu₁₂(cdc)₁₂ building unit. Together, these two adsorption sites form the basis for the lowest occupancy site (0.28(3)) in a triangular pocket between two of the metal-bound CD₄ and a carbazole-bound CD₄ with methane−methane distances of 2.79−4.13 Å (Figure 4). At double the methane loading, a number of additional sites are populated. The occupancy of the first two sites are essentially unchanged while the occupancy of the third site increases to 0.37(3). At this loading, the final two triangular windows of the cage are populated at their three corners. An additional adsorption site on the inside of the cage and the copper on the exterior bind methane at this loading. Increasing the methane dose to 4.0/ Cu reveals no additional CD₄ binding sites, although the overall occupancy increases. Interestingly, at this loading the small cage in PCN-81 remains the only one that is occupied. Furthermore, the concurrent population of multiple adsorption sites in PCN-81 is consistent with the shallow nature of the adsorption isotherm, which indicates a lack of very strong methane binding sites. Eight methane adsorption sites are related by symmetry to afford a total methane capacity of 49 molecules per cage. This corresponds to a methane adsorption capacity of approximately 260 cm³/g, a value PCN-81 reaches by 35 bar. At this pressure, the two additional pores in the material remain unoccupied but would allow for additional storage at higher pressures. At lower pressures, the two other pores are essentially empty space that is detrimental to the density, and thus volumetric capacity of the material. This is consistent with the nearly equal volumetric adsorption capacities of PCN-81, Cr₁₂(cdc)₁₂, and Mo₁₂(cdc)₁₂ at 35 bar and 298 K.

Figure 4.

(Upper) Structure of activated PCN-81. (Lower) CD₄ adsorption in PCN-81 as determined by powder neutron diffraction. (Left) At low loading three sites in the Cu₁₂(cdc)₁₂ portion of the structure are populated. (Right) Up to 4.0 CD₄/Cu, methane exclusively resides in the octahedral cage. With 12 Cu²⁺ ions per cage, this results in a cage capacity of 49 methane molecules.

In conclusion, we have shown that with precise synthesis, solvent exchange, and activation procedures, high surface areas are attainable for porous coordination assemblies. Although the surface areas displayed by these materials currently fall short of the record values displayed by metal−organic frameworks, the materials discussed here show the potential utility of porous coordination cages for high-pressure gas storage. Neutron diffraction experiments suggest the M₁₂(cdc)₁₂ cage that is also present in PCN-81 and a number of related MOFs may be the optimal pore environment for high-pressure methane storage. An ideal methane storage material may consist nearly entirely of M₁₂(cdc)₁₂ cages, whether it is molecular or an extended solid. Future work along these lines will involve tailoring cage stability, solubility, and gas uptake via ligand functionalization.