“Orthogonal-Twisted-Arm” Ligands for The Construction of Metal-Organic Frameworks: New Topology and Catalytic Reactivity

Abstract

Extended tetratopic benzoic acid ligands with “orthogonal-twisted-arms” conformations were designed and synthesized for the construction of new MOF structures (OTA-MOF). Upon coordination with Cd²⁺ and Cu²⁺ cations, two well-defined new MOFs were prepared. X-ray single crystal structures were successfully obtained, demonstrating the formation of a new topology (4,4,4-c). The OTA2-MOF-Cu gave moderate stability in organic solvents and good gas sorption ability toward CO₂. This new MOF showed superior catalytic reactivity toward the epoxide-CO₂ cycloaddition, giving >50 folds yield enhancement over the controlled reaction without MOF. It is expected that this new ligand design, porous structure, and excellent CO₂ catalytic reactivity will make OTA-MOF promising new materials for applications in catalysis and separation.

Graphical Abstract

Extended tetratopic benzoic acid ligands with “orthogonal-twisted-arms” conformations were designed and synthesized for the construction of new MOF structures (OTA-MOF). X-ray single crystal structures were successfully obtained, demonstrating the formation of a new topology (4,4,4-c). The OTA2-MOF-Cu gave moderate stability in organic solvents and good gas sorption ability toward CO₂.

As an important class of porous materials, metal-organic frameworks (MOFs) have gained tremendous attention in the past decades with their high surface area, tunable pore sizes, diverse geometries, and accessible function/reaction sites.^[1] These characters make MOFs promising materials for chemical, material and biological research,^[2] including gas sorption/separation,^[3] chemical catalysis,^[4] molecular sensing^[5] and drug delivery^[6]. It is well known that the overall reactivity/property of these porous materials is depending on the basic building blocks (ligands) and metal cation binding mode (topology).^[7] In this work, we report the design and synthesis of “Orthogonal-Twisted-Arm” (OTA) ligands for the preparation of a new class of porous MOF materials. With two different binding units with orthogonal geometry, OTA ligands coordinate with Cd²⁺ and Cu²⁺ cations to form two new types of OTA-MOFs with structures characterized by X-ray crystallography. Interesting new topologies were revealed with this new ligand design shown in Scheme 1 using extended benzene core structure for MOF construction. Application of OTA2-MOF-Cu as a catalyst for epoxide-CO₂ condensation gave >50 folds enhanced catalytic reactivity over the cupper salt catalysis (No MOF), suggesting the future of this unique ligand system in the construction of new functional porous materials. A combination of ligand binding sites and metal-joint geometry will ultimately determine the shape of resulting porous structures.^[8] Benzoic acid and imidazole have been used as effective binding head for effective syntheses of MOFs and ZIFs.^[9] Another crucial factor in MOF structure design is the ligand shape and control the topology of the resulting porous complexes.^[10] As shown in Scheme 1A, among all the reported multitopic ligands systems, bis-benzene is a popular core due to its superior stability and easy functionalization.^[9b,11] In all di-aryl system, the twisted conformation between the two aryl rings is adopted due to the A-1,3 repulsion.^[12] Many reported systems focused on extension of substitutions, such as 1,3,5-trisubstituted,^[9b] 1,2,4,5-tetra-substituted^[11b] or fully substituted benzenes^[11c]. All these structures are highly symmetrical. The less symmetrical 1,2,3-trisubstituted benzene core has never been reported for MOF construction, likely due to the reduced symmetry ligands with lower stability. This is often due to that MOF networks are synthesized upon the establishment of reaction equilibrium and the symmetric and highly ordered networks are often produced with high overall thermal stability. Despite the challenge, using reduced symmetry linkers for coordinating groups has been reported to approach high performance materials.^[13] Thus, this unique scaffold could be interesting since it might offer two different metal coordination directions both horizontally and vertically, leading to three-dimensional cross-linking. (Scheme 1B) To testify this hypothesis, we design the orthogonal twisted arm ligand OTA1 and OTA2 for MOF synthesis.

Scheme 1.

Extended benzene core structure for MOF construction

We designed OTA ligands with three benzene as the backbone and tetra acid substitution on the two sides of the molecule. The overall structures resemble a human figure with two arms and two legs stretching out on each side as carboxylates. This twisted structure with four branch points enables the possibility to bind to metal ions with various combinations (hand to hand, hand to foot, foot to foot). With the dual-layer design, the two arms and two foot are expected to connect metal cations in both horizontal and vertical directions to enhance the network stability.

A general synthetic route was developed from readily available 4-methyl aniline for large scale ligand preparation. Within 8–10 linear steps, both ligands OTA1 and OTA2 could be successfully prepared (PXRD details in ESI). With the ligands available, we conducted the MOF synthesis upon reaction with metal cations. Fortunately, a stable complex was obtained with OTA1 ligand with Cd²⁺ as shown in Figure 1A. Complex resembled a “Twin ion engine line starfighter” with two inner-layer (foot) carboxylate coordinated to Cd²⁺, giving the extended network. Although this complex network does not give the desired porous material since the two upper layers failed to coordinate with Cd²⁺ cations due to the extended length, the overall structure confirmed our key hypothesis that the OTA scaffold provides orthogonal binding possibility with the twisted bi-aryl conformation. Encouraged by this result, we put our focus on evaluating various MOF synthesis conditions with OTA2 ligand as it has a shorter arm. When treating OTA2 and Cd(NO₃)₂·4H₂O with 1:4 ratio in DMF and H₂O solvents, a transparent crystal was obtained while treating the mixtures at 85 °C after 72 hours (Figure 1B). The FT-IR spectra showed the characteristic band of coordinated carboxylate groups at 1686 and 1413 cm⁻¹ for asymmetric and symmetric stretches (Figure S1 and S2). The broad band at 2866 cm⁻¹ for carboxylic acid stretching disappeared, indicating the resultant binding between carboxylate and the metal. With similar protocols with Cu(NO₃)₂·2.5H₂O at 90 °C temperature for 24 hours were carried out to afford blue crystals, of which the MOF formation were evidenced by FT-IR spectra. (Figure S3).

Figure 1.

A) OTA ligands synthesis and synthesis of OTA-MOFs. B) Cd cluster. C) View of abstracted crystal structure of OTA2-MOF-Cd. along a axis. D) Cu clusters. E) View of abstracted crystal structure of OTA2-MOF-Cu along c axis. The cyan ball was used to estimate the size of the molecule fitting in the pore.

X-ray single crystal structures of these two complexes were successfully determined as OTA2-MOF-Cd and OTA2-MOF-Cu. Notably, with different bonding modes, Cd²⁺ and Cu²⁺ often lead to different coordination patterns. This is the case for these two MOFs formed from the same ligand OTA2. On the other hand, the twisted OTA ligand has four branch points, enabling various binding combination of binding mode (including hand to hand, hand to foot, foot to foot). In OTA2-MOF-Cd, each cluster possesses two 6-coordinated Cd cations with different binding modes. In one case, Cd₁ binds to 2 outer carboxylates (hand to hand) and 1 inner carboxylate. Cd₂ binds with four carboxylates from four ligands (left hand, right hand, left foot, right foot).

In OTA2-MOF-Cu, octahedrally shaped Cu paddle wheel SBUs were obtained with four points of extension (carboxylate C atoms). Interestingly, two different types of linkages were observed: 1) outer-outer (hand to hand) and inner-inner (foot to foot) carboxylates chelate in the diagonal direction, and 2) outer-inner (hand to foot) and inner-outer (foot to hand) carboxylates chelate in the diagonal direction. The overall network consists of hexagon repeating units with methyl group pointing into the center with an ideal pore size of 12 Å. The underlying 3D network can be described as a (4,4,4)-c net with the point symbol [(4.6³.8²) (4².6².8²) (4³.6³)₂] as calculated by ToposPro software ^[14a] (Figure S4–S9). The total void volume in OTA2-MOF-Cu is 57.6% after eliminating guest and coordinated water using the PLATON program ^[14b].

To explore the property and functionality of these two new networks, we first evaluated the stability of OTA2-MOF-Cd and OTA2-MOF-Cu under various conditions. PXRD patterns of both samples were first collected to confirm their phase purity. The diffraction pattern of both OTA2-MOF-Cu and the calculated data from its X-ray single crystal structure was coherent (see ESI). A series of PXRD experiments were carried out using the pre-treated MOF samples by soaking in a variety of organic solvents, including MeOH, ACN, DCM, and THF. Using solvents to deal with OTA2-MOF-Cd presented a destructive breakdown of porous frameworks manifested by almost no signals observed in the PXRD spectra. In contrast, the signals after treatment in these organic solvents remained similar to the calculated spectrum, implying moderate stability of OTA2-MOF-Cu after being immersed in organic solution for 24 h. (Figure S13) The difference shown in organic stability may be owing to the more rigid coordination conformation, highly organized structure of OTA2-MOF-Cu, which enables it to maintain its framework. It is also likely that the hydrophobic pore invites the organic solvents to come in and hold the whole framework. In contrast, the lack of porosity and flexible binding in Cd cluster caused the structure to collapse upon guest molecule exchange.

Thermogravimetric analysis was also performed. MOFs showed excellent stability up to 400 °C in the air. The OTA2-MOF-Cd and OTA2-MOF-Cu samples maintained their structures until 400 °C and 386 °C respectively. Both samples reached complete collapse at a similar temperature around 380 ºC (Figure S14 and S15).

To investigate the gas sorption capacity of the synthesized MOF materials, N₂ adsorption data were collected at 77 K to examine the porosity of activated OTA2-MOF-Cu samples. (Figure 2A) The adsorption plot of OTA2-MOF-Cu shows a reversible type-I isotherm which is characteristic for porous materials with a steep increase at P/P₀ = 0.1. It gave the maximum N₂ uptake of 208 cm³ g⁻¹ at P/P₀ = 0.95. The results were employed to estimate their surface areas. Fitting the adsorption isotherms of CO₂ to Brunauer-Emmett Teller (BET) equation gave the surface areas of ~715 m² g⁻¹ (Langmuir surface area: ~817m² g⁻¹). Furthermore, a total pore volume of 0.322 cm³ g⁻¹ was estimated using the Dubinin-Radushkevich equation (Table S1).

Figure 2.

A) N₂ sorption of OTA2-MOF-Cu at 77 K. B). FT-IR patterns of OTA2-MOF-Cu. C) CO₂ sorption isotherm of OTA2-MOF-Cu at 273 K and 298 K. D) Sorption selectivity of CO₂ over methane selectivity plotted based on IAST calculations.

Carbon dioxide (CO₂) capture and utilization (CCU) is of importance or the environmental concern with growing atmospheric CO₂ emissions caused by the use of fossil fuels.^[15] As a chemically stable porous material, we envision OTA2-MOF-Cu to have the ability to capture and store CO₂ for conversion and functionalization. To evaluate CO₂ uptake performances of OTA2-MOF-Cu, CO₂ adsorption isotherms were collected on the activated samples at 273 K and 298 K, as shown in Figure 2C. The results revealed 22.26 cm³ g⁻¹ and 10.76 cm³ g⁻¹ maximum adsorption of CO₂ at 273 K and 298 K for OTA2-MOF-Cu respectively. The reported value of CO2 absorption of HKUST-1 (10.60 cm³/g, 298K) ^[16] is similar to that of OTA2-MOF-Cu in this work. This may due to the similar pore size of the two MOFs (1.0 Å and 1.2 Å respectively). The fair adsorption ability could be ascribed to the hydrophobic core in the framework. The value of CO₂/CH₄ selectivity of 2 was obtained by ideal solution adsorbed theory (IAST) with good correlation factor. (Figure 2D).

With the excellent stability, CO₂ adsorption capacity, we set out to apply this OTA2-MOF-Cu for the cycloaddition of CO₂ to epoxide reaction and the catalytic performance was evaluated. The percent yields were determined by GC (see ESI for detailed calculation). As shown in Table 1 entry 1, the activated OTA2-MOF-Cu achieved a complete conversion from epoxide to cyclic carbonate 50 times higher yield compared to entry with no addition of the catalyst. The optimal MOF loading could be as low as 0.65 mol% per exposed copper site. (Table S2). In contrast, under the same condition, both CuBr and CuBr₂ could only give 10% of yield. These results were also compared with the literature reported MOFs with Cu paddle wheel cluster such as HKUST-1. (Table S2) HKUST-1 demonstrated moderate reactivity of 60% in the transformation of propylene epoxide. The high reactivity of OTA2-MOF-Cu is presumably due to the open coordination cite and fast diffusion to the hydrophobic pore. The yield maintained as 94% in the case of vinyl glycidyl ether, implying that the high porosity of OTA2-MOF-Cu facilitated the reaction. With further increase of the steric hindrance of R substituent, the yield decreased greatly to 58% as shown in Table 1, entry 7 for phenol glycidyl ether. Notably, the OTA2-MOF-Cu could be recycled without much significant decrease in the catalytic performance after five cycles. (Figure S21) These results revealed that the performance of chemical fixation of CO₂ for cycloaddition reaction can be significantly improved by twisted conformational under ambient condition, where the hydrophobic core trap and concentrate substrate for faster reaction and the open Cu(II) coordination site could possibly serve as the Lewis acid to activate the epoxide.

Table 1.

Catalytic CO₂-epoxide condensation with OTA2-MOF-Cu^[a]

entry	R	[Cu] cat.	Yield[b]
1	CH3	MOF-Cu	100%
2	CH3	no cat.	2%
3	CH3	CuBr2	10%
4	CH3	HKUST-1	60%
5	C2H5	MOF-Cu	97%
6	CH2=CHCH2OCH2	MOF-Cu	94%
7	C6H5OCH2	MOF-Cu	58%

^[a]general conditions: substituted oxirane (2 mmol), 0.65 mol% per exposed copper site, TBABr (0.2 mmol), CO₂ (1 atm), neat, rt.

^[b]yield was determined by GC.

In summary, OTA2-MOF-Cd and OTA2-MOF-Cu (with new topology) were developed for the first time using tetratopic benzoic acid with extended branch points as the linker. OTA2-MOF-Cu exhibited superior stability compared to OTA2-MOF-Cd likely due to the rigid Cu paddle wheel coordination mode and hydrophobic pore. These unique features provided on the potential catalytic site for simple using MOF as the catalyst in Lewis acid promoted catalysis without breaking the MOF frameworks. Owing to the good chemical stability, it has significantly promoted the CO₂ cycloaddition reaction by trapping the CO₂ gas and epoxide in the hydrophobic core while activate the epoxide ring opening. Further extending the side branch of this core structure to incorporate the new network into material and biological applications is currently undergoing in our lab.