- Yadav, Ankit K.
- Gładysiak, Andrzej
- Song, Ah-Young
- Gan, Lei
- Simons, Casey R.
- Alghoraibi, Nawal M.
- Alahmed, Ammar H.
- Younes, Mourad
- Reimer, Jeffrey A.
- Huang, Hongliang
- Giner Planas, José
- Stylianou, Kyriakos C.
The capture of carbon dioxide (CO2) is crucial for reducing greenhouse emissions and achieving net-zero emission goals. Metal-organic frameworks (MOFs) present a promising solution for carbon capture due to their structural adaptability, tunability, porosity, and pore modification. In this research, we explored the use of a copper (Cu(II))-based MOF called m CBMOF-1. After activation, m CBMOF-1 generates one-dimensional channels with square cross sections, featuring sets of four Cu(II) open metal sites spaced by 6.042 Å, allowing strong interactions with coordinating molecules. To investigate this capability, m CBMOF-1 was exposed to ammonia (NH3) gas, resulting in hysteretic NH3 isotherms indicative of strong interactions between Cu(II) and NH3. At 150 mbar and 298 K, the NH3-loaded (∼1 mmol/g) material exhibited a 106% increase in CO2 uptake compared to that of the pristine m CBMOF-1. Carbon-13 solid-state nuclear magnetic resonance spectra and density functional theory calculations confirmed that the sequential loading of NH3 followed by CO2 adsorption generated a copper-carbamic acid complex within the pores of m CBMOF-1. Our study highlights the effectiveness of sequential pore functionalization in MOFs as an attractive strategy for enhancing the interactions of MOFs with small molecules such as CO2., Introduction
Carbon dioxide (CO2) emissions into the atmosphere contribute to the greenhouse effect (1,2) caused by the absorption of infrared rays reflected by the Earth’s surface. (3,4) Uncontrolled greenhouse gas emissions into the atmosphere have played a pivotal role in global climate change. (5,6) According to the data from the Mauna Loa Observatory, (6) the atmospheric concentration of CO2 reached 422.1 ppm in July 2023, marking a 31.9% increase from the 1950 level of 320 ppm and a 50.8% increase from preindustrial levels of 280 ppm. (7,8) Despite efforts to set a target of 350 ppm for global atmospheric CO2 levels, CO2 concentrations continue to rise. (9,10) In 2018, the Intergovernmental Panel on Climate Change (11) reported that surpassing a global temperature increase of 1.5 °C above the 20th-century average could lead to severe consequences, and crossing the 2.0 °C threshold might result in irreversible outcomes. Moreover, higher global temperatures are expected to exacerbate issues such food and water scarcity, potentially leading to increased poverty rates. (12,13) Carbon dioxide capture presents a promising solution to reduce the amount of CO2 entering the atmosphere or remove CO2 already emitted, such as through the use of Direct Air Capture systems, which were considered a viable option in many outlook scenarios and are essential in the IEA net-zero pathway. However, CO2 capture technologies face challenges due to high costs associated with material limitations such as working capacity, degradation, or extensive energy demand. Therefore, the development of innovative approaches and new materials is crucial to capture of CO2 effectively and economically.
Metal–organic frameworks (MOFs) are a class of porous materials composed of metal ions or clusters and organic linkers, which form extended structures. Their properties, which include exceptionally high surface areas, (14) tailored functionality, (15,16) and permanent porosity, (17,18) make them highly suitable for gas adsorption applications such as CO2 capture. (19,20) Several approaches have been explored to enhance the interaction between the pores of MOF adsorbents and CO2. (21) One such approach involves the utilization of microporous (pore size <20 Å) and ultramicroporous (pore size <7 Å) MOFs, whose pores possess favorable shapes and sizes to facilitate host–guest interactions (11,22−24) effective for capturing CO2 even in the presence of water vapor. (25) In a recent study, MIL-120, an ultramicroporous MOF, was employed for CO2 capture, resulting in enhanced CO2 uptake under humid conditions. (26) Another well-explored method of CO2 capture enhancement involves the introduction of open metal sites through rational design, desolvation, or activation. Many MOFs, including those from the M-MOF-74 family (M = Mg, Mn, Fe, Co, Zn, Ni), (27,28) zirconium-based UiO-66, (29) and HKUST-1, (30) have achieved significant CO2 uptakes using this approach.
Through careful design, either before or after their synthesis, the affinity of MOFs toward CO2 can be significantly improved. (31−36) One key approach involves pore functionalization, wherein terminal uncoordinated functional groups such as −NH2, −OH, and −COOH are introduced into the organic linker before MOF synthesis. (33,37−39) This results in MOFs with pores decorated with polarizable functional groups, thereby leading to increased CO2 uptakes. In a study conducted by Arstad et al., the impact of an amino-substituted 1,4-benzene dicarboxylate ligand on the CO2 uptake of three different MOFs (USO-1-Al, USO-2-Ni, and USO-3-In) was investigated. (40) Interestingly, in all three of the MOFs, the amino-substituted MOF exhibited higher CO2 uptakes than their unfunctionalized counterparts. USO-2-Ni-NH2 demonstrated the highest CO2 uptake of 14 wt % (3.18 mmol/g) at 1 bar and 298 K, which represented a discernible increase from 10 wt % (2.27 mmol/g) for the unfunctionalized USO-2-Ni at the same conditions. (40) Other examples of such functionalized MOFs include CAU-1, (41) bio-MOF-11, (42) and NH2-MIL-53(Al). (43)
The pore surface in MOFs can be modified postsynthesis. (33,37,44) Examples of such postsynthetic MOF functionalization include attaching alkylamines to open metal sites, which changes the CO2 capture mechanism and significantly increases CO2 uptake. Darunte et al. found that after loading the MIL-101(Cr) MOF with tris(2-aminoethyl)amine, the CO2 uptake increased from ∼0.5 mmol/g in pristine MOF to ∼1.4 mmol/g in amine functionalized MOF at 150 mbar and 298 K. (45) Lyu et al. studied the amino acid functionalized MOF-808 for CO2 capture in humid flue gas conditions, and they found that after loading glycine in the MOF, the CO2 uptake increased from ∼0.2 mmol/g in the MOF-808 to ∼0.5 mmol/g in MOF-808-Gly at 150 mbar and 298 K. (46) Finally, the functionalization of the pores of Mg2(dobpdc) with N,N′-dimethylethylenediamine resulted in about a 10-fold increase in the CO2 uptake. (47−49) In this class of MOFs, CO2 is not directly attached to open metal sites, replacing alkylamine molecules, but is instead cooperatively inserted in between the metal and the amine.
In our study, we expand on the postsynthetic functionalization approach by coordinating gaseous ammonia (NH3) to the open metal sites of a Cu(II)-based MOF, mCBMOF-1, to enhance CO2 adsorption. The activated MOF features four closely positioned, coordinatively unsaturated Cu(II) centers, which act as potent Lewis acidic sites known for their strong affinity for Lewis bases like NH3. By anchoring NH3 molecules within the MOF pores, we introduce a Lewis basic environment that enhances the selective capture of the Lewis acidic CO2 molecules. This stepwise coordination of NH3, followed by the coordination of CO2, leads to the in situ formation of carbamic acid within the MOF structure. Our approach is unique in leveraging the internal pore chemistry to enhance CO2 capture, demonstrating that the incorporation of a Lewis basic species within a MOF can significantly amplify its CO2 adsorption capacity.
Results and Discussion
Synthesis and Characterization
The MOF employed in our study is a copper paddlewheel-based MOF known as mCBMOF-1, with the chemical formula [Cu2(mCB-L)2(DABCO)0.5(H2O)]·guest molecules, in which mCB-L refers to 1,7-di(4-carboxyphenyl)-1,7-dicarba-closo-dodecaborane and DABCO represents 1,4-diazabicyclo[2.2.2]octane (Figure 1). (50,51) The synthesis of mCBMOF-1 is carried out in a mixture of water, ethanol, and dimethylformamide (DMF) (in a ratio of 1:5:5) at 80 °C for 48 h. Powder X-ray diffraction (PXRD, Figure 2a) was utilized to verify the identity and purity of the material, while Fourier transform infrared (FTIR) spectra (Figure S1) confirmed the successful incorporation of the ligands into the MOF structure. The MOF crystallizes in tetragonal space group I422, and the atomic coordinates allow for the following chemical interpretation of the structure. The mCB ligands, which feature a chevron shape with an opening angle of 117°, are linked by Cu2 paddlewheels, forming a 44 net (sql). The top and the bottom of this two-dimensional layer are decorated with aqua and DABCO ligands, which complete the coordination sphere of Cu2 paddlewheels. The DABCO ligands connect the two-dimensional layer to neighboring ones, and two such layers are related to each other by a 21-axis (rotation by 180° followed by translation by b/2). Because DABCO ligands are present on both sides of each two-dimensional layer, they extend the coordination to a three-dimensional 44·66 net (sqp). The structure of mCBMOF-1 consists of two interpenetrating nets, which collectively occupy 67.6% of the unit-cell volume, leaving 32.4% as void space, as determined by the Void routine of the Olex2 crystallographic software. (52) The void system comprises one-dimensional channels running along [110], interconnected along [001] by narrow segments (Figure 1d). Removal of aqua ligands from the crystal structure increases the void volume to 34.5% and generates an additional channel along [001] (Figure 1e, Figure S2), which is potentially accessible for gas diffusion. The channel is lined with four Cu(II) open metal sites situated at the vertices of a square, with a distance of 6.042 Å between the adjacent Cu(II) atoms and 8.545 Å between opposite ones (Figure S3). This four-Cu2 site plays a key role in the adsorption properties of mCBMOF-1 (see below).
Figure 1
Figure 1. Structure representation of mCBMOF-1. (a) mCBMOF-1 comprises a meta-carborane-dicarboxylate ligand and DABCO. (b) In mCBMOF-1, each Cu(II) is five coordinated, forming Cu2-paddlewheels. The axial position for each Cu in the paddlewheel is distinct: Cu1 is bound to an aqua ligand, and Cu2 is bound to the N atom of DABCO. (c) The extended structure of mCBMOF-1 illustrates the generation of the 4Cu site where each Cu is connected to an aqua ligand. (d) Three-dimensional ball and stick packing of mCBMOF-1. Red surfaces represent the accessible void within the unit cell. (e) The pore topology of mCBMOF-1 after removal of the four aqua ligands bound to the 4Cu site. This results in the formation of an additional void channel extending along [001]. Atom color code: gray for C, red for O, blue for N, golden for B, sky blue for Cu, and pale yellow for H.
Figure 2
Figure 2. Solid-state characterization and sorption isotherms. (a) The powder X-ray diffraction patterns of as-made mCBMOF-1 (shown in red) are in excellent agreement with the pattern derived from the crystal structure of this material (simulated; shown in black), confirming the bulk phase purity of the MOF. The PXRD pattern of the ∼1 mmol NH3-loaded mCBMOF-1 (shown in blue) indicates that the MOF retains its crystallinity. (b) The N2 isotherm for the activated mCBMOF-1, obtained at 77 K and 1 bar, indicates its microporous nature. (c) CO2 isotherms for the activated mCBMOF-1 collected at 298 and 303 K. (d) Ammonia isotherms of mCBMOF-1. The second NH3 isotherm (colored in blue) was measured after mCBMOF-1 was reactivated following the first isotherm (colored in red). Filled symbols: adsorption; empty symbols: desorption.
Thermogravimetric analysis (TGA) of mCBMOF-1 shows a 19% weight loss attributed to the removal of solvent molecules (water, DMF) present in the voids of the mCBMOF-1 (Figure S4). The thermal activation of mCBMOF-1 induces a color change from sky blue to deep navy (Figure S5), which is consistent with the removal of coordinated water molecules and the presence of open Cu2+ sites. (53) When the activated mCBMOF-1 is exposed to air, it returns to its original sky-blue color due to the recoordination of water molecules present in air to active Cu2+ centers. This phenomenon is in line with observations seen in other Cu(II)-based MOFs such as HKUST-1 (54,55) and Cu-MOF-2. (53) Furthermore, the ultraviolet–visible (UV–vis) spectrum of mCBMOF-1 exhibits an absorption peak with a λmax of 744 nm (Figure S6). This absorption peak corresponds to the d–d electronic transition in Cu2+ and gives the material a sky-blue color.
Upon exposure of the activated mCBMOF-1 to gaseous NH3 at 1 bar for >2 h, it undergoes a color change to dark purple (Figure S7), and its UV–vis spectrum displays an absorption peak at λmax of 647 nm (Figure S6). These changes in the optical properties arise from the coordination of NH3 to open Cu(II) sites and the positioning of NH3 in the spectrochemical series. A comparison with the pristine mCBMOF-1 (λmax = 744 nm) provides valuable insights: given that H2O is a weak field ligand and NH3 is a strong field ligand, the energy required for the d–d transition in an amine complex is higher compared to the aqua complex. (56) The PXRD pattern of NH3-loaded mCBMOF-1 (Figure S8) reveals that the MOF becomes amorphous, indicating that the metal–ligand bonds in the MOF are disrupted. FTIR spectroscopy shows an additional peak at 3320 cm–1, corresponding to N–H stretching in the bound NH3 molecule (Figures S2 and S9). When mCBMOF-1 was exposed to a low NH3 concentration (∼1 mmol/g loading), its PXRD pattern remained crystalline. The TGA profile of this material is comparable to that of the as-made MOF (Figure S4), indicating that NH3 loading and binding to Cu sites do not affect the material’s stability. Furthermore, 1H and 13C solution NMR in DMSO-d6 of the controlled (∼1 mmol/g) NH3-loaded mCBMOF-1 demonstrates the stability of the MOF with no leaching of the mCB ligand in the solution (Figures S10 and S11).
Adsorption Properties
At 77 K and 1 bar, activated mCBMOF-1 adsorbs N2, as indicated by its type I adsorption isotherm (Figure 2b, Figure S12a), with a BET surface area of 996 m2/g. The single point adsorption total pore volume at p/p0 = 0.90 is 0.516 cm3/g, which is consistent with 0.632 cm3/g derived from the static crystal structure of mCBMOF-1 (refined against data collected at 100 K). The activated mCBMOF-1 displays a low affinity for CO2, evident from the quasi-linear shape of its adsorption isotherm (Figure 2c) and a relatively low isosteric heat of CO2 adsorption (Qst) of 28 kJ/mol (Figure 3b). This Qst value positions mCBMOF-1 among the MOFs with moderate performance toward CO2 capture such as HKUST-1 (Qst = 24–28 kJ/mol), (30,57) UiO-66 (Qst = 26 kJ/mol), (29) and MIL-100(Fe) (Qst = 30 kJ/mol) (57) but falls behind Mg-MOF-74 (Qst = 47 kJ/mol) with its open metal sites. (27)
Figure 3
Figure 3. Carbon dioxide uptake and isosteric heat of adsorption. (a) Comparison of the CO2 isotherms for activated mCBMOF-1 (shown in red) and NH3-loaded (shown in green) at 298 K. At 150 mbar (orange arrow), the NH3-loaded (∼1 mmol/g) mCBMOF-1 exhibited 106% higher CO2 uptake compared to that of mCBMOF-1. (b) Isosteric heat of CO2 adsorption calculated for the activated (red) and NH3-loaded (green) mCBMOF-1..
To investigate the NH3 adsorption properties of mCBMOF-1, NH3 isotherms were recorded (Figure 2d). Pure (99.995%) NH3 isotherms collected at 298 K and 1 bar demonstrated that 1 g of activated mCBMOF-1 uptakes 11.5 mmol of NH3. This uptake corresponds to 43.6 molecules of NH3 per unit cell (Figure 2d), a quantity comparable to what is observed in other MOFs featuring open metal sites, suggesting a strong affinity for NH3. (58−63) This remarkable uptake can be explained by two consecutive phenomena. At low pressures, NH3 (Lewis base) forms coordination bonds with open Cu sites (Lewis acid), allowing a maximum of four NH3 molecules to bind in one unit cell of activated mCBMOF-1. As the pressure increases, NH3 fills the pores through dispersive interactions and H-bonds. The high affinity of activated mCBMOF-1 for NH3 is further supported by the hysteresis observed in its isotherm (Figure 2d) and the density functional theory- (DFT-) calculated energy of −119.8 kJ/mol for the preferred NH3 adsorption site (Figure S13). The high value of adsorption energy indicates that NH3 undergoes chemical bonding within the MOF, implying that the unsaturated Cu sites act as the primary adsorption sites. (60) Additionally, the desorption curves reveal that initially only a small amount of NH3 is released, and even as the pressure decreases to zero, a substantial amount of NH3 remains in the MOF structure and pores. During the second isotherm, the hysteresis observed in the NH3 isotherm at 298 K indicates a maximum uptake of mCBMOF-1 at 5.9 mmol/g (22.5 molecules/unit cell; Figure 2d). The lower NH3 uptake during the second NH3 isotherm is likely attributed to some NH3 molecules being strongly bound to its active sites during the first cycle. However, the BET surface area of the NH3-loaded mCBMOF-1 is 6 m2/g, suggesting that NH3 molecules disrupt its long-range order, and it becomes nonporous to N2 at 77 K and 1 bar (Figure S12b). When the amorphous NH3-loaded material is immersed in water, the resulting material becomes crystalline (Figure S8). This material is likely to be a new structure, as the Bragg reflections do not match those of mCBMOF-1, and is porous to N2 at 77 K (Figure S12c). Efforts to elucidate the structure of this new phase through single-crystal X-ray diffraction have been unsuccessful due to the polycrystalline nature of the material. Saturating mCBMOF-1 with NH3 renders it nonporous to N2 at 77 K, as NH3 molecules occupy both the Cu(II) coordination sites and all the pores. To prevent this, we exposed the activated mCBMOF-1 to a controlled amount of NH3 gas, thereby achieving an uptake of ∼1.0 mmol/g, equivalent to one NH3 molecule binding to one open Cu site. The resulting material (∼1.0 mmolNH3/g loading) was porous to N2 at 77 K and 1 bar, and its BET surface area was 686 m2/g (Figure S12d). The BET surface area of NH3-loaded (∼1.0 mmol/g) mCBMOF-1 is lower than the activated mCBMOF-1, likely due to a partial loss of crystallinity and/or occupation of the pores with NH3. Interestingly, controlled ammonia-loaded (∼1 mmol/g) mCBMOF-1 does not collapse and can be fully regenerated upon immersion in water, as confirmed by PXRD (Figure S14a). The BET surface area of regenerated mCBMOF-1 is 923 m2/g (Figure S14b). Its regeneration is thought to be due to the recoordination of water molecules to the 4Cu2 sites, displacing NH3 molecules from the MOF structure. (56) These observations confirm the stability of mCBMOF-1 when it is loaded with ∼1 mmol/g of NH3. Elemental analysis of the NH3-loaded (∼1 mmol/g) mCBMOF-1 indicated the presence of 1.9 molecules of NH3 per formula unit, [Cu2(mCB-L)2(DABCO)0.5(NH3)1.9–x]·xNH3·0.3H2O (Table S1), some of which are coordinated to the open Cu centers of mCBMOF-1. (64) Overrepresentation of the NH3 molecules in this formula may stem from the inherent imprecision of CHN analysis.
Mechanism of CO2 Adsorption
The mCBMOF-1 loaded with a controlled amount of adsorbed NH3 was further investigated toward CO2 capture. As shown in Figure 3a and Figures S15 and S16, the material features type I bent-shaped CO2 adsorption isotherms, a distinctive feature evident when compared to the CO2 isotherm of activated mCBMOF-1 (Figure 2c). The consistency in the shape of the CO2 isotherms was verified across three distinct experiments (Figure S16), emphasizing the reliability and reproducibility of our findings. At 150 mbar and 298 K, conditions relevant to postcombustion carbon capture, the activated MOF exhibited a CO2 uptake of 0.51 mmol/g. However, after doping activated mCBMOF-1 with a controlled amount of gaseous NH3, the uptake at the same pressure and temperature increased to 1.05 mmol/g (Figure 3a). A similar increase was observed in Qst at low coverage, which was 28 kJ/mol for the activated MOF and increased to 39 kJ/mol for the NH3-loaded mCBMOF-1 (Figure 3b). These findings indicate different CO2 adsorption mechanisms in the two discussed materials. The enhanced adsorption can be attributed to the synergy of acid–base interactions and hydrogen bonding of CO2 and the NH3 groups bound to the MOF pores through the open metal sites. (40,65−68) The CO2 isotherm of the NH3-loaded mCBMOF-1 does not exhibit an adsorption hysteresis, a characteristic that has not been detected in other chemisorption-based materials as well. (49,69)
FTIR and solid-state NMR techniques were employed to investigate the interactions between CO2 and NH3 within the pores of mCBMOF-1. In the FTIR spectra, new peaks emerged at 3300 and 480 cm–1 upon NH3 loading into the activated MOF, corresponding to N–H and Cu–N stretching frequencies, (70) respectively (Figure S9). In the NH3–CO2 loaded mCBMOF-1, the appearance of additional peaks at 3366 and 1633 cm–1 indicates coordinated NH3 vibrations, (71−73) suggesting the possible formation of carbamic acid upon sequential introduction of NH3 and CO2 into the pores of mCBMOF-1.
Solid-state 13C NMR was employed to investigate the structural modification occurring upon the introduction of NH3 and subsequent CO2 loading in the paramagnetic mCBMOF-1. (74−76) This method was adapted to explore the binding of NH3 and CO2 in mCBMOF-1. Figure 4 illustrates the 13C solid-state NMR spectra for the as-made and NH3-CO2 loaded mCBMOF-1. In the spectrum of the as-made mCBMOF-1 (Figure 4a), the resonances observed at 31.1, 34.6, and 164.0 ppm are attributed to the carbon atoms of the guest dimethylformamide (DMF) molecules located within the MOF pores. The resonance at 49.6 ppm corresponds to the carbon atom of the m-carborane group, while the peaks within the range of 132 to 150 ppm represent the aromatic C of the benzoate fragment of the mCB ligand. Additionally, the peak at 173.3 ppm corresponds to carbonyl C of the carboxylate group of the ligand. The resonance at 79.7 ppm indicates a single environment for the carbon atoms in the coordinated DABCO. This resonance is shifted to the downfield region due to paramagnetic shift, compared to the free DABCO molecule, which appears at 47.9 ppm (Figure S17). Some aspects of the NMR spectra present paramagnetic shifts that are the subject of further study. The ss-NMR of the activated MOF shows that the peaks at 31.1, 34.6, and 164 ppm are absent, confirming the removal of DMF molecules from the pores of the MOF (Figure S18a). All other peaks in the as-made MOF remain in the activated MOF, confirming its structural stability. The successful loading of NH3 in activated mCBMOF-1 is evidenced by two distinct resonances at 79.7 and 72.0 ppm for the DABCO ligand in the MOF structure. These resonances suggest two distinct environments for the carbon in DABCO: DABCO–Cu-dimer–NH3 and DABCO–Cu-dimer–open Cu site (Figure S18b). The broadened resonance at 169.9 ppm, corresponding to the carboxylic carbon of the mCB ligand, can be attributed to a dynamic environment around the Cu-dimer in NH3-loaded mCBMOF-1. Other peaks closely match the position and intensity of those in activated mCBMOF-1, underscoring the chemical similarity of the two materials.
Figure 4
Figure 4. Solid-state 13C NMR spectra for (a) as-made mCBMOF-1 (inset: a fraction of mCBMOF-1 and corresponding assignments for its individual carbon atoms). The resonances at 31.1, 34.6, and 164.0 ppm correspond to the guest DMF molecules encapsulated within the MOF pores. (b) CO2 loading on the NH3-functionalized mCBMOF-1. The resonances corresponding to DMF are absent, confirming its successful activation followed by NH3 and CO2 loading. The resonance at 161.6 ppm corresponds to the carbonyl carbon of carbamic acid, indicating that CO2 is chemisorbed upon loading on the NH3-loaded mCBMOF-1.
The successful loading of CO2 in the NH3 loaded mCBMOF-1 is evident from the more distinct peaks in the 13C NMR spectrum compared to those of the as-made MOF (Figure 4b, Figure S18c). In the NH3-CO2 loaded MOF, the resonances at 81.6 and 75.8 ppm are attributed to DABCO. The presence of multiple resonances for DABCO carbon atoms indicates the emergence of distinct chemical environments within DABCO. Given that DABCO is bound to one side of the Cu(II) paddlewheels, occupying the opposite coordination site can change the electronic structure of Cu(II) paddlewheels, as evidenced by the chemical shifts of DABCO. Importantly, the presence of two nonequivalent sites, which are not equally populated, suggests that only a portion of Cu sites are coordinated with the NH3 molecules. This is further supported by the observed peak splitting at 44.8 and 57.1 ppm, which is thought to be due to the distortion of the mCB ligand in the MOF structure. Our findings are consistent with our elemental analysis revealing that each formula unit of mCBMOF-1 hosts 1.9 molecules of NH3 (chemisorbed and/or physisorbed). This observation has significant implications for our DFT results described below. The resonances at 172.9 and 169.8 ppm corresponding to carboxylic carbon of the mCB ligand can be attributed to a dynamic environment around the Cu(II) paddlewheels in the NH3-CO2 loaded mCBMOF-1. (74) The resonances ranging from 122 to 150 ppm correspond to the aromatic carbons of the mCB ligand, while the peak at 44.8 ppm corresponds to m-C of the carborane core in the ligand. In line with the literature, the peak at 161.6 ppm corresponds to carbon in carbamic acid (H2N–COOH). (47,49,77,78) This peak differs significantly from the carbon peak at 166.1 ppm in ammonium carbamate (Figure S19), suggesting the formation of carbamic acid within the pores of our MOF.
The DFT sheds light on the interactions between Cu(II)-paddlewheels and CO2 or NH3 via isosurface plots of charge differences (Figure S20) and the formation of carbamic acid within the pores of the MOF structure (Figure 5). For the coadsorption intermediate state, the NH3 molecule is adsorbed on an open Cu(II) site (Figure 5b,f(I)), while CO2 molecules interact with the adjacent open Cu(II) site (Figure 5c,f(II)). The coadsorbed CO2 and NH3 molecules further interact with each other to generate a motif (Cu(II)–NH3–CO2), which is an exothermic step with the calculated reaction energy of −15.2 kJ/mol (Figure 5d,f(III) and Figure S13). At the transition state (TS), the H atom of NH3 approaches the O atom of the CO2, forming the O═C---N–H intermediate. The calculated energy of activation is 18.8 kJ/mol, indicating that the formation of −NH2COOH is kinetically feasible (Figure 5e,f(IV)) with a further energy release of 23.8 kJ/mol. These results clearly indicate that the presence of the coordinated NH3 molecule, anchored in the unique environment of four open Cu(II) sites located in proximity to one another, significantly improves the immobilization of CO2. This enhancement results in the formation of the Cu(II)–NH3–CO2 intermediate followed by the stable Cu(II)–NH2COOH complex within the pores of mCBMOF-1.
Figure 5
Figure 5. Mechanistic overview of the CO2 capture with mCBMOF-1. DFT optimized configurations of (a) activated mCBMOF-1, (b) NH3 absorbed in activated mCBMOF-1, (c) CO2 adsorbed in NH3-loaded mCBMOF-1, (d) transition state of CO2 interacting with NH3 within the pore of mCBMOF-1, (e) product of the CO2 + NH3 reaction within the pore of mCBMOF-1, and (f) a schematic illustration of the carbamic acid formation pathway, steps I–IV. Atom color code: gray for C, red for O, blue for N, sky blue for Cu and pale yellow for H.
The pathway described in our work for the formation of carbamic acid differs significantly from the mechanism reported with Mg2-(dobpdc) and diamines. (49,79−81) Mg2-(dobpdc) is a mesoporous MOF with a pore size of 22 Å having six Mg(II) open metal sites in the pore. (82) When diamine is introduced into the MOF, it interacts with the metal in the pore, forming coordination bonds. Subsequently, CO2 is cooperatively inserted into the Mg–N bond, forming a carbamate in dry conditions and carbamic acid if water is present. (47,81) The landscape in our microporous mCBMOF-1 is unique, and both NH3 and CO2 coordinate to metal centers first and then, due to the proximity of those centers, react to form carbamic acid (Figure 5f).
Conclusions
Our study reports, for the first time, that sequential pore functionalization in a MOF with open Cu(II) sites leads to a 106% enhancement of CO2 adsorption at low pressures. The activated mCBMOF-1 comprises repeating units with four active open Cu(II) sites positioned in proximity to one another on vertices of a square cross section of the pore; upon exposure to NH3, some of these Cu(II) sites, acting as a Lewis acid, interact with molecules of NH3, a Lewis base, through coordination bonding. The remaining open Cu(II) sites are available for subsequent CO2 coordination. When both CO2 and NH3 are anchored in the unique coordination environment of the pore, they interact with each other, forming a carbamic acid species, as inferred from our FTIR, 13C solid-state NMR, and DFT results. This interaction is evident as NH3-functionalized pores attract CO2 more strongly than the nonfunctionalized activated mCBMOF-1: the CO2 uptake at 150 mbar and 298 K increases by 106%, and Qst increases by 40%. Discovering MOFs with unique pores geometries and active sites and applying postsynthetic functionalization could pave the way for further advancements in CO2 capture from dilute sources.
Experimental Section
Materials
All the materials and chemicals used in this study were bought from commercial sources such as MilliporeSigma, Sigma-Aldrich, and Tokyo Chemical Industry and used without further purification.
Synthesis of mCB-H2L
1,7-Di(4-carboxyphenyl)-1,7-dicarba-closo-dodecaborane ligand (mCB-H2L) was synthesized as per the literature procedure. (50)
Synthesis of mCBMOF-1 [(Cu2(mCB-L)2(DABCO)0.5(H2O))·2DMF·2H2O]
mCBMOF-1 was synthesized as per the reported procedure. (50) Briefly, DABCO (6.5 mg, 0.059 mmol), mCB-H2L (90 mg, 0.234 mmol), H2O (1 mL), and DMF (5 mL) were mixed in an 8 dram vial, and the mixture was sonicated until all the solids were dissolved. Next, Cu(NO3)2·6H2O (68 mg, 0.234 mmol) was added in ethanol (5 mL) and sonicated until the metal salt was completely dissolved. The two solutions were mixed and heated at 80 °C for 48 h to obtain green crystals. Finally, the crystals were washed with DMF and acetone and dried at 80 °C to obtain mCBMOF-1.
Activation of mCBMOF-1
A sample of mCBMOF-1 was enclosed in a Schlenk tube and heated under a vacuum for 24 h at 120 °C using an oil bath to remove water molecules from the pores.
Loading mCBMOF-1 with Excess NH3
Ammonia loading was conducted using either the 3FLEX gas analyzer from Micromeritics or the Schlenk methods. In the Schlenk method, the knob of the tube containing activated mCBMOF-1 was charged with pure ammonia gas (utilizing a pressure gauge) at 1 bar for ∼1 h, sealed, and allowed to equilibrate for an additional ∼1 h.
Regeneration of mCBMOF-1 after Ammonia Loading
Ammonia loaded MOF was soaked in water for 15 min with stirring. After that, it was washed with water followed by acetone and dried at 80 °C.
Loading mCBMOF-1 with a Controlled Amount of NH3
Pristine mCBMOF-1 was introduced into a borosilicate glass tube and activated under a vacuum for 24 h at 120 °C. A set amount adsorbed of ∼1.00 mmolNH3/g was achieved with a 3FLEX Adsorption Analyzer.
Loading the NH3-loaded mCBMOF-1 with CO2
mCBMOF-1 loaded with ∼1.00 mmolNH3/g was transferred to a Parr reactor, which was then filled with CO2 gas to 1 bar at room temperature and ∼1 h to achieve equilibrium.
Data Availability
The data that support the findings of the study are included in the main text and the Supporting Information file. Raw data can be obtained from the corresponding author upon request. Solid-state NMR raw data collected from 400 MHz have been deposited on Dryad at https://doi.org/10.5061/dryad.v15dv426b. (83)
Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.4c00808.
Detailed characterization methods (PXRDs, TGA, FTIR, gas sorption, ss-NMR, dynamic CO2 capture), characterization data, and computational studies (PDF)
au4c00808_si_001.pdf (8.22 MB)
Sequential Pore Functionalization in MOFs for Enhanced Carbon Dioxide Capture
21
views
0
shares
0
downloads
Skip to figshare navigation
S1Supporting InformationSequential Pore Functionalization in MOFs for Enhanced Carbon Dioxide CaptureAnkit K. Yadav,aAndrzej Gładysiak,aAh-Young Song,bLei Gan,c,dCasey R. Simons,eNawal M. Alghoraibi,fAmmar H. Alahmed,fMourad Younes,fJeffrey A. Reimer,bHongliang Huang,g*José G. Planas,cKyriakos C. Stylianoua*aMaterialsDiscovery Laboratory (MaD Lab), Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, United StatesEmail: kyriakos.stylianou@oregonstate.edubDepartment of Chemical and Biomolecular Engineering, University of California, Berkeley 94720, USA cInstitut de Ciència de Materials de Barcelona (ICMAB-CSIC), Bellaterra 08193, SpaindSchool of Chemistry and Materials Science, Nanjing Normal University, Nanjing 210023, P. R. ChinaEmail: huanghongliang@tiangong.edu.cneCenter for Advanced Materials Characterization in Oregon, University of Oregon, 1443 E, 13th AVE, Eugene, Oregon 97403, USAfARAMCO, R-GC 335, Floor 3, Research and Development Center (Building 2297), Dhahran, Saudi ArabiagStateKey Laboratory of Separation Membranes and Membrane Processes, School of Chemistry and Chemical Engineering, Tiangong University, Tianjin 300387, China
S2ExperimentalMethodsPowder X-Ray Diffraction (PXRD):The materials' bulk phase purity and crystallinity were validated using PXRD recorded on a Rigaku SmartLab X-ray diffractometer (monochromated Cu Kαradiation,λ= 1.5418Å). At ambient temperature, measurements were taken for the 2θ range of 3 -30degrees with astep increment of 0.01 degrees and a speed of 0.75°/min. Voltage and current were measured to be 40kV and 44 mA, respectively.Thermogravimetric analysis (TGA): Shimadzu Thermogravimetric Analyzer was used to confirm the thermal stability of the materials. Temperatures were measured from 25°C to 600°C at a heating rate of 10°C/min in an argon environment with a flow rate of 100 mL/min.Fourier-transform infrared spectroscopy (FT-IR):The Perkin Elmer Spectrum Two FTIR Spectrometer was used for all FT-IR measurements.Gas Sorption Measurements: Nitrogen adsorption-desorption isotherms were measured at 77 K using a3FLEX Adsorption Analyzer. Before the measurement, the samples were activated at 120 °C for 12 hoursunder vacuum. The BET surface areas were determined using the BET (Brunauer-Emmett-Teller) model. At 298 K, 303 K, and 308 K, CO2adsorption-desorption isotherms were measured.Isosteric heat of CO2adsorption (Qst) was determined using Clausius-Clapeyron equation from the data recorded at 298 K and 303 K.The 3FLEX adsorption analyzer was also used for themeasurement ofNH3isothermsat 298 K.Solid-state NMR experiments.13C solid-state NMR spectra were collected on Bruker Avance NEO (1H Larmor frequency 400 MHz) using a 3.2 mm probe. The 13C spectra were acquired under Magic Angle Spinning (MAS) at a spinning rate of 20 kHz. A spin echo sequence was employed without decoupling. For the CO2-NH3-loaded mCBMOF-1, the spectrum was obtained from an average of 725,472 scans with a recovery delay of 100ms. The NH3-loaded mCBMOF-1spectrum was averaged over 917,504 scans with a recovery delay of 50ms. For the as-made mCBMOF-1, the data consisted of an average of 458,752 scans with a recovery delay of 50ms. The 13C isotropic chemical shifts were referenced to adamantane at 38.48 ppm.DFT Calculations:All DFT calculations were carried out using the CP2K code.1All calculations employed a mixed Gaussian and planewave basis sets. Core electrons were represented with norm-conserving Goedecker-Teter-Hutter pseudopotentials,2-4and the valence electron wavefunction
Share
Download
figshare
Terms & Conditions
Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.
Author Information
Corresponding Authors
Hongliang Huang - State Key Laboratory of Separation Membranes and Membrane Processes, School of Chemistry and Chemical Engineering, Tiangong University, Tianjin 300387, China; Orcidhttps://orcid.org/0000-0001-9690-9259; Email: huanghongliang@tiangong.edu.cn
Kyriakos C. Stylianou - Materials Discovery Laboratory (MaD Lab), Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, United States; Orcidhttps://orcid.org/0000-0003-1670-0020; Email: kyriakos.stylianou@oregonstate.edu
Authors
Ankit K. Yadav - Materials Discovery Laboratory (MaD Lab), Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, United States
Andrzej Gładysiak - Materials Discovery Laboratory (MaD Lab), Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, United States; Orcidhttps://orcid.org/0000-0003-3302-1644
Ah-Young Song - Department of Chemical and Biomolecular Engineering, University of California, Berkeley 94720, United States
Lei Gan - Institut de Ciència de Materials de Barcelona (ICMAB-CSIC), Bellaterra 08193, Spain; School of Chemistry and Materials Science, Nanjing Normal University, Nanjing 210023, P. R. China
Casey R. Simons - Center for Advanced Materials Characterization in Oregon, University of Oregon, 1443 E, 13th Ave, Eugene, Oregon 97403, United States; Orcidhttps://orcid.org/0000-0002-8024-6996
Nawal M. Alghoraibi - ARAMCO, R-GC 335, Floor 3, Research and Development Center (Building 2297), Dhahran 31311, Saudi Arabia
Ammar H. Alahmed - ARAMCO, R-GC 335, Floor 3, Research and Development Center (Building 2297), Dhahran 31311, Saudi Arabia
Mourad Younes - ARAMCO, R-GC 335, Floor 3, Research and Development Center (Building 2297), Dhahran 31311, Saudi Arabia
Jeffrey A. Reimer - Department of Chemical and Biomolecular Engineering, University of California, Berkeley 94720, United States
José G. Planas - Institut de Ciència de Materials de Barcelona (ICMAB-CSIC), Bellaterra 08193, Spain; Orcidhttps://orcid.org/0000-0002-1648-2169
Author Contributions
CRediT: Ankit Yadav conceptualization, formal analysis, investigation, methodology, visualization, writing - original draft; Andrzej Gładysiak conceptualization, formal analysis, investigation, methodology, visualization, writing - original draft; Ah-Young Song formal analysis, investigation, methodology, writing - review & editing; Lei Gan investigation, writing - review & editing; Casey Simons investigation, writing - review & editing; Nawal M. Alghoraibi methodology, writing - review & editing; Ammar H. Alahmed methodology, writing - review & editing; Mourad Younes methodology, writing - review & editing; Jeffrey A. Reimer conceptualization, funding acquisition, methodology, supervision, writing - review & editing; Hongliang Huang conceptualization, funding acquisition, investigation, supervision, writing - review & editing; Jose Giner Planas conceptualization, funding acquisition, methodology, supervision, writing - review & editing; Kyriakos C. Stylianou conceptualization, funding acquisition, investigation, methodology, supervision, visualization, writing - original draft.
Notes
The authors declare no competing financial interest.
Acknowledgments
K.C.S., A.G., and A.K.Y. thank Dr. Patrick Reardon for assistance with the NMR experiments. MaD Lab thanks Marilyn and Brian Kleiner for their generous support of this project through their donor advised fund. K.C.S. acknowledges support from Saudi Aramco for the development of this project and thanks Dr. Aqil Jamal for useful discussions. K.C.S. thanks the Department of Chemistry at Oregon State University (OSU) for support through start-up funding and College of Science Industrial Partnership Award. A.K.Y. acknowledges support from the Department of Chemistry for the Milton Harris fellowship (2023). K.C.S. thanks Baydin Inc., maker of the Boomerang productivity suite, for supporting this work. J.G.P. thanks Spanish Ministerio de Ciencia e Innovación project PID2022-136892NB-I00 and the Generalitat de Catalunya (2021/SGR/00442) for financial support. A.S. gratefully acknowledges support as a Pines Magnetic Resonance Center Postdoctoral Fellow. We thank Drs. Hasan Celik, Raynald Giovine, and UC Berkeley’s NMR facility in the College of Chemistry (CoC-NMR) for spectroscopic assistance. The instrument used in this work is supported by the National Science Foundation under Grant No. 2018784., With funding from the Spanish government through the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2023-001263-S)., Peer reviewed