Hexagonal Layer Manganese Metal-Organic Framework for Photocatalytic CO2 Cycloaddition Reaction.

A novel manganese metal-organic framework (Mn-MOF) termed UAEU-50 assembled from a benzenedicarboxylate linker (BDC) and trinuclear manganese clusters was synthesized and fully characterized using different spectroscopic and analytic techniques (e.g., X-ray powder diffraction, UV-vis diffuse reflectance spectroscopy, thermogravimetric analysis, scanning electron microscopy, and energy-dispersive X-ray spectroscopy). UAEU-50 adopted a hexagonal layer structure and exhibited superior thermal stability and robust chemical stability. Photocatalytic activities of UAEU-50 were investigated using the cycloaddition of CO2 to different epoxides, forming cyclic carbonates. Impressively, UAEU-50 can transform up to 90% photocatalytic CO2 conversion to cyclic carbonates in the visible-light region at ambient conditions.

Introduction

A cheap and renewable carbon dioxide (CO2) can serve as a C1 feedstock due to its availability, nontoxicity, and nonflammability.[1] Researchers have been looking for methods to recycle it back into stable and useful products.[2,3] Among these, CO2 cycloaddition to epoxides is one of the most lucrative routes because of the atom economy efficiency of the process which is important for sustainable development, for example to produce more value-added compounds.[2,4] One of these is cyclic carbonates which are used as electrolytes for Li-ion batteries,[5] as polar aprotic solvents,[6] as monomers in polycarbonate synthesis,[7] and as intermediate compounds widely applied in the biomedical and pharmaceutical industries.[8] Additionally, cyclic carbonates are essential in the synthesis of disinfectants and herbicides (as fuel additives).[9,10]

Traditionally, cyclic carbonate formation from cycloaddition of CO2 and epoxides is achieved hydrothermally using catalytic systems under an energy-intensive process that requires high temperature and pressure.[11] Recently, a wide range of catalytic systems[12] have been developed; nonetheless, there are challenges of such an energy-demanding reaction that yields poor catalytic efficiency and poor recyclability. Therefore, many attempts at using photocatalytic materials have been implemented.[1]

To date, several metal–organic frameworks (MOFs) have been hydrothermally utilized for CO2 cycloaddition reaction with epoxides.[13,14] Based solely on the Lewis acidity of the metal center, it would appear that the most active catalysts would be hard acid metal centers such as lanthanides[15] and early transition-metal-based MOF catalysts.[16] As an example, hard Lewis acidic lanthanide-based MOFs, reported by Thammakan et al.,[15] possessed vacant coordination sites with 2,2′-dinitrobiphenyl-4,4′-dicarboxylate and showed around 79% conversion of epoxide to cyclic carbonates through CO2 cycloaddition reaction. The CO2 cycloaddition with epichlorohydrin could also achieve a nearly quantitative conversion yield when using a series of lanthanide MOFs based on 3,3′,5,5′-azobenzenetetracarboxylic acid.[17]

Borderline Lewis acid transition metallic nodes, on the other hand, have also been given high conversion yield, such as the isostructural series of M-MOFs (M = Mg, Co, Ni, and Zn) composed of a 4,4′-(ethyne-1,2-diyl)bis(2-oxidobenzoate) linker.[18] Among these frameworks, Zn-MOFs have exhibited the highest catalytic activity of CO2 cycloaddition with a percent yield of 96%.[18] Nonetheless, MOFs constructed from borderline or soft metal nodes have also been used and exhibit similar performance for other CO2 to cyclic carbonate formation reactions such as a functionalized Pd–Eu-MOF nanocatalyst,[19] which achieved around 98% yield of cyclic carbonates. The Cu3(BTC)2 catalyst reported by Macias et al.[20] exhibited moderate conversion yields for chloropropene carbonate from cycloaddition of CO2 and epichlorohydrin.

Indeed, MOFs have drawn much interest due to their unique properties based mainly on but not limited to the well-defined structures, high density of active catalytic sites, and porosity. These properties render MOFs useful in many applications including storage,[21,22] separation,[23,24] sensing,[25,26] catalysis,[27,28] biomedical applications,[29,30] and many more.[31,32] Recently, our group has studied the applicability of MOFs and MOF composites for photocatalyzing the cycloaddition of carbon dioxide to epoxides, leading to various cyclic carbonates.[33,34]

Many MOFs suffer from low thermal and/or photochemical stability, which significantly hinders their application for solar-to-chemical conversion such as water splitting and CO2 reduction and/or utilization.[35,36] In this regard, MOFs built from one of the most earth-abundant metals (i.e., Mn) have shown efficient photo-/electrocatalytic activities toward CO2 reduction.[37] As also considered to be a borderline Lewis acid, Mn was often coordinated in frameworks with different geometry suitable for the CO2 cycloaddition reaction. For example, a series of structurally controlled Mn-MOFs for catalytic CO2 cycloaddition reaction made from N-phenyl-N′-phenylbicyclo[2.2.2]oct-7-ene-2,3,5,6-tetracarboxdiimide dicarboxylic acid showed high yields with different epoxides.[38]

Herein, we report the synthesis and characterization of a new Mn-MOF based on trinuclear Mn clusters and benzenedicarboxylate (BDC) moieties, termed UAEU-50 (UAEU stands for United Arab Emirates University). Especially, UAEU-50 adopting a 2D structure with a hexagonal layer net (hxl) presents excellent photocatalytic cycloaddition reaction of CO2 to epoxides under visible-light irradiation at mild synthetic conditions, which is one of the first examples of Mn-MOFs used for the CO2 cycloaddition purpose.

Results and Discussion

UAEU-50 was synthesized under modified experimental conditions reported by Ladrak et al.[39] Especially, dimethylformamide is replaced with dimethylacetamide (DMA). In a typical procedure, H2BDC (99 mg, 0.5475 mmol) and Mn(NO3)2·4H2O (92 mg, 0.365 mmol) were dissolved in 10 mL of DMA and then transferred to a 75 mL Pyrex sealed tube. The sealed tube was then placed in a preheated oven at 120 °C for 24 h. Then, the white crystalline product was isolated by filtration, washed with acetone, and dried at 50 °C under a vacuum for 24 h.

The crystal structure of UAEU-50 (Figure ) was determined by SC-XRD. UAEU-50 was crystallized in a triclinic crystal system with unit cell parameters of a = 16.42 Å, b = 16.93 Å, and c = 18.28 Å and angles of α = 72.69°, β = 69.02°, and γ = 75.10°. UAEU-50 was composed of trinuclear Mn clusters connecting with BDC linking units. The Mn central atom of the trinuclear cluster adopts six coordinates from six carboxylates, forming a perfect octahedra. This central Mn atom links to the other 2 Mn atoms. Each of them connects to (i) 4 O atoms from 3 carboxylates and (ii) to 2 O atoms from DMA guest molecules. These two Mn atoms adopt the distorted octahedral geometry. Based on the topological perspective, Mn3 clusters act as points of extension, which are linked together by six linear BDC units. Therefore, we can simplify Mn3 clusters to be 6-coordinate units, and the combination of these planar 6-coordinate nodes form hxl topology (Figure ). UAEU-50 possesses very small trigonal windows (3.5 Å, Figure ). The phase purity of UAEU-50 was confirmed by PXRD analysis (see Supporting Information, SI, Figure S1). Upon comparing the diffraction peaks of the synthesized MOF with the simulated ones obtained from SC-XRD, some peaks were not present, which could be due to preferred orientation (as observed by SEM images, UAEU-50 crystallized in long and thin needle crystals), which is common for 2D MOFs.[40−42]

Figure 1

Benzenedicarboxylate (BDC) reacted with trinuclear Mn clusters to form UAEU-50, presenting a hexagonal layer structure with a pore size of 3.5 Å. It is worth noting that hxl-a stands for an augmented net of hxl topology. Atom color: C, gray; O, red; N, blue; Mn, cyan. Hydrogen atoms and DMA guest molecules in UAEU-50s crystal structure are omitted for clarity.

Especially, UAEU-50 exhibited homogeneous needle-like crystals as proven by SEM analysis (SI, Figures S2 and S3). EDX analysis of the prepared UAEU-50 confirmed the presence and expected percentage of C, O, N, and Mn atoms (SI, Figure S3, Table S1).

The band gap energy of UAEU-50 was determined by plotting photon energy (hν) vs (αhν) as shown in Figure . The Tauc method was used considering that the electron transitions in UAEU-50 are direct allowed transitions. Data from the DRS spectrum and Tauc method were used to calculate the band gap energy, 3.04 eV, which represented a light absorption in the visible region.

Figure 2

UAEU-50 possesses a band gap energy of 3.04 eV estimated based on UV–vis DRS using the Tauc plot.

UAEU-50 was thermally analyzed using a thermogravimetric technique under a nitrogen atmosphere with a heating rate of 10 °C min–1 (SI, Figure S4). Between 130 and 200 °C, 14% weight of the sample was lost, attributed to the trapped solvent molecules. The major weight loss (49%) was observed between 500 and 550 °C, corresponding to the exothermic loss of the organic linker and the formation of manganese oxide which decomposed at higher temperatures.

Upon performing the photocatalysis, the chemical stability of UAEU-50 was tested by immersing 10 mg of the activated sample in 5 mL of several solvents including methanol, acetone, acetonitrile, and ethanol at room temperature for 24 h. Comparing the PXRD pattern of each sample to the activated MOF, it was evident that UAEU-50 exhibits superior stability in all above-mentioned solvents (SI, Figure S5). It is worth noting that the intensity of the peaks in the UAEU-50s PXRD pattern changed in acetonitrile. This could be due to the fact that acetonitrile has smaller size compared to the others that entered the pores and thus changed the relative intensity of the peaks in the PXRD pattern of UAEU-50.

UAEU-50 is a promising candidate for photocatalytic reactions based on its band gap value of 3.04 eV and the high stability in organic solvents. To study UAEU-50’s photocatalytic activity, we carried out the photocatalytic cycloaddition of CO2 and epoxides to cyclic carbonates (Scheme ). The photocatalytic reaction was accomplished in a Pyrex sealed tube, where 0.28 mmol of styrene oxide was mixed with a binary cocatalyst (tetrabutyl ammonium bromide, Bu4NBr, 9 mg) and UAEU-50 (10 mg) in a mixture of 1:5 mL of methanol/acetonitrile. The reaction was placed on a magnetic stirrer and irradiated with a 400 W halogen lamp for 24 h. The product was obtained by filtration, followed by extraction with dichloromethane. Other products were obtained from different epoxides following the same procedure.

Scheme 1

Synthesis of Different Cyclic Carbonates from Epoxides and CO2

Substrate 1,2-epoxy-2-methylpropane (1c) has no Ha attached to the same carbon, accordingly; calculation based on the account of the other hydrogen attached to the adjacent carbon.

According to eq , conversion of each reaction was calculated by comparing the 1H NMR integrals of OCH protons in the starting material (1Ha) and the relevant proton in the product (1Hb) (SI, Figures S6–S11).[43]

Table lists the chemical shifts of the protons of the epoxide substrates and the corresponding cyclic carbonate products along with their percent yields. Based on the obtained results, moderate to high photocatalytic activity of UAEU-50 was observed. The highest conversion yield was 90% for the cycloaddition of 1,2-epoxy-2-methylpropane to CO2, forming an isobutylene carbonate, while the rest of the reaction yields were in the range between 20% and 54%. It is tempting to conclude that aliphatic epoxide (Table , entry 1c) with less steric hindrance produces the highest yield.

Table 1

Chemical Shift of OCH Protons in Each Epoxide and Carbonate and Their Relative Cyclic Carbonate Yielda

epoxide	δOCH (CDCl3) (epoxide, 1Ha)	δOCH (CDCl3) (carbonate, 1Hb)	product yieldb (%)	TONc	TOF (h–1)d
1a	3.83	5.65	54%	32	1.3
1b	3.83	5.65	20%	12	0.5
1c	2.46	4.12	90%	54	2.2
1d	2.87	4.5	39%	23	1
1e	2.86	4.52	27%	16	0.7
1f	3.04	5.10	24%	14	0.6

Reaction conditions: epoxide (1.429 mmol), photocatalyst (10 mg, 0.024 mmol), Bu4NBr (9 mg, 0.028 mmol), and 0.045 mmol of carbon dioxide at 353 K and 24 h.

Yield of isolated product was determined by 1H NMR spectroscopy.

TON = (mmol of product)/(mmol of catalyst).

TOF = (mmol of product)/(mmol of catalyst) (reaction time, hour).

To prove the reaction is mainly photocatalytic in nature, control experiments were conducted using 1,2-epoxy-2-methylpropane as a substrate since it gave the highest yield of 90% (Table ). The conversion yields were low in the absence of a Bu4NBr cocatalyst or the use of UAEU-50 photocatalyst alone; the percent yield was found to be 12% and 20%, respectively (entries II and III, Table ). This indicated the need for the photocatalyst along with the cocatalyst for the reaction to proceed. When the reaction was carried out in the dark at room temperature in all three cases, in the presence of both photocatalyst and cocatalyst (entry IV), only photocatalyst (entry V), and only cocatalyst (entry VI), yields were very low (8%, 3%, and 0%, respectively), indicating that this reaction requires a leverage that is induced by visible light. Even though when the reaction was done in the dark and heated at 80 °C (entry VII), only 17% conversion was obtained. In entry VIII, using light, commercial TiO2 was used as a photocatalyst along with Bu4NBr that yielded 30%, while 4% was obtained in the presence of TiO2 alone without the cocatalyst (entry IX) which is very low compared to UAEU-50 reaction yields. When the reaction was carried out in the absence of the photocatalyst UAEU-50 and the cocatalyst, no conversion yield was observed (entry X). Comparing these results with the typical reaction conditions (entry I) showed the importance of the presence of photocatalyst and cocatalyst to complete the reaction with high yield.

Table 2

Control Experiment of Cycloaddition of 1,2-Epoxy-2-methylpropane to Isobutylene Carbonate

entry	photocatalyst	yieldb %	TONc	TOF (h–1)d
I	UAEU-50, nBu4NBr, light	90%	54	2.2
II	nBu4NBr, no UAEU-50, light	12%	7	0.3
III	UAEU-50, no nBu4NBr, light	20%	12	0.5
IV	UAEU-50, nBu4NBr, no light	8%	4.8	0.2
V	UAEU-50, no light	3%	1.8	0.9
VI	nBu4NBr, no UAEU-50, no light	0%	-	-
VII	UAEU-50, nBu4NBr, no light, heat (353 K)	17%	10	0.4
VIII	TiO2, nBu4NBr, light	30%	15	0.6
IX	TiO2, no nBu4NBr light	4%	2.4	0.1
X	no UAEU-50, no nBu4NBr, light	0%	-	-

Reaction conditions: epoxide (1.429 mmol), photocatalyst (10 mg, 0.024 mmol), Bu4NBr (9 mg, 0.028 mmol), and 0.045 mmol of carbon dioxide at 353 K and 24 h.

Yield of isolated product was determined by 1H NMR spectroscopy.

TON = (mmol of product)/(mmol of catalyst).

TOF = (mmol of product)/(mmol of catalyst) (reaction time, hour).

A possible mechanism for the CO2 cycloaddition can be postulated as follows: the Lewis acidic (Mn nodes) sites of UAEU-50 coordinated with an oxygen atom from the epoxides result in weakening of the C–O bond of the epoxide. After that, the epoxide ring is opened by the cocatalyst TBAB attack on the C atom.[44] At this stage, CO2 is activated by the photogenerated ligand-to-metal charge-transfer process under continuous light irradiation. Finally, the activated CO2 attacks C in epoxides, liberating the corresponding cyclic carbonates (Scheme ). Due to the fact that the BDC linker is not a strong visible-light absorption antenna, the photoexcited transformation from the linker to metal may not be straightforwardly advanced. Therefore, we think that the active Mn sites in UAEU-50 played an important role in activating the cycloaddition reaction. Exhaustive efforts have tried to replace BDC with BDC-NH2 to even enhance the photocatalytic activity of the MOF but were unsuccessful. However, future investigation for synthesizing Mn-MOFs based on visible-light-driven linkers is highly encouraged.

Scheme 2

Proposed Mechanism of Cyclic Carbonate Formation from Epoxides and CO2

Nonetheless, light is believed to play an important role in this process: when UAEU-50 with energy larger than 3.04 eV is exposed to light (energy gap between HOMO and LUMO orbitals of UAEU-50), electrons and holes are generated. CO2 activation is initiated through coordination of the Mn ions to the UAEU-50 framework, and photogenerated electrons further facilitate cycloaddition reaction. Moreover, photogenerated holes of the UAEU-50 framework would promote the ring opening of epoxide.

To obtain more insight into the reusability of UAEU-50, cycling experiments were conducted using the cycloaddition reaction of CO2 to 1,2-epoxy-2-methylpropane. The reaction was done under the same conditions as described earlier. The photocatalyst was separated from the reaction by centrifugation and then rinsed with methanol to remove any trace impurities from the reaction mixture. The solid was then dried in an oven at 80 °C for 4 h. The amount of solid recovered after each cycle was approximately 95% of the original amount that we start with, and this solid was utilized for the cycling test. Especially, PXRD patterns clearly showed that UAEU-50 retained its crystalline structure after four consecutive reaction cycles (SI, Figure S6). More importantly, after four consecutive runs there is no significant loss in activity observed, as indicated by 1H NMR spectra (SI, Figure S14).

Conclusion

We have successfully synthesized a novel new Mn-based MOF, termed UAEU-50, bearing an hxl layer topology. UAEU-50 exhibited superior thermal and chemical stability. Furthermore, UAEU-50 showed a band gap energy of 3.04 eV, making it a potential candidate for photocatalytic reaction in the visible-light region. As a model photocatalytic cycloaddition reaction of CO2 to epoxides, UAEU-50 exhibited moderate to high conversion rate to cyclic carbonate formation.