Construction of solvent-dependent self-assembled porous Ni(II)-coordinated frameworks as effective catalysts for chemical transformation of CO₂

Abstract

Coordinated frameworks having large pores can accommodate organic molecules and adsorb gas, such as waste carbon dioxide, to transform into valuable chemical products. By incorporating a tetraphenylethylene moiety as the four-point connected node, herein, two types of Ni(II)-based coordinated frameworks have been solvothermally synthesized via solvent driven self-assembly and structurally characterized. Notably, these two complexes both contain a trinuclear cluster but with different styles and construction modes. The IR and XRD spectra suggested that the two compounds are highly stable in various solvents, including water, DMF, C₂H₅OH, C₂H₂Cl₂, DMSO and 1,4-dioxane. Catalysis experiments indicated that both of them could act as effective heterogeneous catalysts for cycloaddition of CO₂ with epoxides, forming cyclic carbonates under ambient conditions. Notably, the two catalysts contained the same amount of catalytic sites, but the activity of Ni-1 was significantly lower than that of Ni-2, demonstrating a size-selective catalytic performance.

---

Introduction

From the viewpoint of crystal engineering, the design and self-assembly of versatile coordination frameworks has attracted a great deal of attention over the past two decades.^1–3 Considered to be promising analogues of porous materials, the construction of metal–organic frameworks (MOFs) by various organic linkers and metal ions or clusters has been of special interest because of their intriguing structural diversity and outstanding physical and chemical properties that can be used in numerous fields.^4–6 Carboxylate ligands represent one of the most popular ligand classes in the assembly of MOFs, mainly due to their excellent stability and diverse bridging modes, which can lead to abundant topological architectures with different size and shape of pores.^7–9 However, the self-assembly procedure remains inscrutable, and it is difficult to accurately tailor their structures. As a consequence, the fabrication of carboxylate-based MOFs requires relatively extreme synthesis conditions for the strict control of the resulting structures and there still remains large room to explore.^10–12

As we know, porous MOF materials exhibit remarkable properties for gas storage and adsorption, which make them useful for solving the environmental problem of increasing CO₂ emissions.^13–15 In terms of catalysis and green chemistry, gaseous waste CO₂ is an ideal and abundant C1 source due to its characteristics of safety, non-toxicity, renewability and ready availability;^16–18 thus, the utilization and transformation of CO₂ into valuable and usable chemical products show attractive prospects for the satisfaction of atom economy.^19–21 In this respect, the cycloaddition of carbon dioxide with epoxides to produce cyclic carbonates is quite promising, as the latter compounds are widely used in industry and the incorporation of CO₂ to fabricate these chemicals does not result in any side products.^22–24 More efficient conversion under mild conditions from CO₂ to cyclic carbonates is an ongoing challenge, although this conversion has been widely studied.^25–28

MOF materials, incorporating active metal sites and designable ligands, display the combined advantages of homogeneous and heterogeneous catalyst; for example, high activity and selectivity, as well as recovery and high controllability.^29–31 From a catalytic perspective, MOF catalysts based on permanently porous or channel frameworks are able to utilize minimum amount of metal ions and organic ligands to build maximum surface areas with predictable, controllable, tailorable and post modifiable pores and cavities.^32–34 We herein report the synthesis and catalytic properties of two Ni(II)-based porous frameworks for the cycloaddition of CO₂ by the incorporation of a tetrakis(4-carboxyphenyl)ethylene (H₄TCPE) moiety as the four-point connected node (Scheme 1). These two compounds have been solvothermally synthesized and exhibit diverse structures via solvent driven self-assembly. We envisioned that the partially twisted ethyl core and multiple rotational phenyl rings of the H₄TCPE moiety, together with its diverse coordination modes, would cause different MOF configurations. Both the two structures contain a trinuclear cluster to provide an identical amount of open metal sites during catalysis.

Scheme 1 Presentation of the H₄TCPE ligand and the assembly of Ni(II) ions in different solvents, resulting in isolation of two types of trinuclear clusters and construction frameworks.

Experimental section

Materials and methods

All chemicals were of reagent grade, were obtained from commercial sources, and were used without further purification. All epoxides were purchased from Acros and distilled under a nitrogen atmosphere from CaH₂ prior to use. Carbon dioxide (99.995%) was purchased from Dalian Institute of Special Gases and used as received. The elemental analyses of C, H and N were performed on a Vario EL III elemental analyzer. Inductively coupled plasma (ICP) analysis was performed on a Jarrel-AshJ-A1100 spectrometer. The powder XRD diffractograms were obtained on a Rigaku D/MAX-2400 X-ray diffractometer with Cu sealed tube (λ = 1.54178 Å). IR spectra were obtained as KBr pellets on a NEXUS instrument. Thermogravimetric analyses (TGA) were carried out at a ramp rate of 10 °C min⁻¹ in a nitrogen flow with a SDTQ600 instrument. Gas sorption isotherms were measured using an Autosorb-IQ-C analyzer of Quantachrome. The N₂ and CO₂ adsorption isotherms for desolvated compounds were collected in a relative pressure range from 10 to 1.0 × 10⁵ Pa. ¹H NMR spectra were obtained on a Varian INOVA-400 MHz type spectrometer. Their peak frequencies were referenced versus an internal standard (TMS) at 0 ppm for ¹H NMR.

Preparation

All reagents were used as purchased without further purification. Tetrakis(4-carboxyphenyl)ethylene (H₄TCPE) was prepared according to the literature methods³⁵ and was characterized by ¹H NMR.

Synthesis of Ni-1. A mixture of H₄TCPE (30 mg, 0.06 mmol), Ni(NO₃)₂·6H₂O (100 mg, 0.34 mmol) and L-Pro (13 mg, 0.11 mmol) was dissolved in 4 mL of a solvent mixture composed of DMF and H₂O (2 : 1, v/v) in a screw-capped vial. The resulting mixture was kept in an oven at 100 °C for 3 days. After cooling the autoclave to room temperature, green block single crystals were separated, washed with water and air-dried. Yield: 60% (based on H₄TCPE). Anal. calcd for C₁₅H₈Ni_0.75O_6.75: C 52.95, H 2.37, Ni 12.94%. Found: C 52.35, H 2.03, Ni 12.61%. IR (KBr): 3450 (br, s), 1638 (s), 1542 (m), 1403 (s), 1187 (m), 1108 (w), 1053 (m), 777 (m) cm⁻¹.

Synthesis of Ni-2. The synthesis process for Ni-2 was similar to that of Ni-1, except the mixed solvent were C₂H₅OH and 1,4-dioxane (2 : 1, v/v). Yield: 32% (based on H₄TCPE). Anal. calcd for C_3.75H_2.25Ni_0.25O_1.34: C 53.99, H 2.72, Ni 17.59%. Found: C 54.06, H 2.49, Ni 17.32%. IR (KBr): 3434 (br, s), 1605 (s), 1547 (w), 1404 (vs), 1204 (s), 1106 (w), 1018 (w), 789 (m), 771 (m) cm⁻¹.

Crystallography

X-Ray intensity data were measured on a Bruker SMART APEX CCD-based diffractometer (Mo-Kα radiation, λ = 0.71073 Å) using the SMART and SAINT programs.^36,37 The crystal data were solved by direct methods and further refined by full-matrix least-squares refinements on F² using the SHELXL-97 software and an absorption correction was performed using the SADABS program.³⁸ The remaining atoms were found from successive full-matrix least-squares refinements on F² and Fourier syntheses. Non-H atoms were refined with anisotropic displacement parameters. The hydrogen atoms within the ligand backbones were fixed geometrically at calculated distances and allowed to ride on the parent non-hydrogen atoms. No satisfactory disorder model could be achieved, and therefore the SQUEEZE program implemented in PLATON was used to remove these electron densities.³⁹ Crystallographic data for Ni-1 and Ni-2 are summarized in Table S1.†

Catalysis details

Before the catalysis experiments, solvent exchange experiments were carried out to activate the catalysts by removing the coordinated water molecules and solvent. The crystal samples were immersed in ethanol for 2 days, and then fully dried out at 100 °C under high vacuum for 6 h to obtain the activated samples. TGA was used to demonstrate the activation process. The mol number of the catalysts (30 μmol) was calculated based on the Ni sites in each catalyst after being activated. For catalyst Ni-1 (18.2 mg, 60 μmol), the molecular weight of the asymmetric unit after removal of solvent and coordinated water (C₁₅H₈Ni_0.75O_4.5) was 304.24 g mol⁻¹, with 0.5 catalytically active Ni ions, equivalent to 30 μmol. For catalyst Ni-2 (9.5 mg, 120 μmol), the molecular weight of the asymmetric unit after removal of solvent and coordinated water (C_3.75H₂Ni_0.25O_1.083) was 79.05 g mol⁻¹, with 0.25 catalytically active Ni ions, equivalent to 30 μmol.

In a typical catalytic reaction, a 10 mL Schlenk tube was purged with 1 atm CO₂ using a balloon under solvent free environment. The vessel was set in an oil bath with frequent stirring at a temperature of 313 K for 40 h. At the end of the reaction, the reactor was placed in an ice bath for 20 min and then opened. The catalysts were separated by centrifugation. Then, 1,3,5-trimethylbenzene (120 mg, 1 mmol) was added as an internal standard for the analysis of the product yields by ¹H NMR. The catalysts were collected after each cycle, washed with CH₂Cl₂ and dried by vacuum. The remaining mixture was degassed and fractionally distilled under reduced pressure or purified by column chromatography on silica gel to obtain cyclic carbonate.

Results and discussion

Synthetic effects

The solvothermal reaction of Ni(NO₃)₂·6H₂O and H₄TCPE at 100 °C for 3 days produced stick-shaped crystals Ni-1 in the mixture of DMF and H₂O (v/v = 2 : 1), and block-shaped crystals Ni-2 in the mixture of DMF and CH₃CN (v/v = 2 : 1). The only difference in the synthesis of the isolated crystals was the mixed solvents. As we know, solvent plays an essential role in the coordination capability because of its steric effect, polarity and solubility, leading to the synthesis of supramolecular polymers with diverse connectivity and architectures. Changing the reaction temperature or ratio of raw materials cannot obtain our target compounds, and the accurate modulation of solvent is the key factor to influence the formation of Ni-1 and Ni-2. Elemental and powder X-ray diffraction (XRD) analyses indicate that the bulk samples consist of a pure, single phase.

Structure of Ni-1. Single-crystal X-ray diffraction studies reveal that Ni-1 crystallizes in the orthorhombic with space group Pmna. As shown in Fig. 1a, each asymmetric unit consists of two independent Ni(II) ions connected by a μ₂-O atom, half deprotonated TCPE ligand and one coordinated H₂O molecule. The two Ni atoms are both in six-coordinated distorted octahedral geometries: two O atoms from carboxyl groups of four TCPE ligands, one μ₂-O atom and three O atoms from H₂O molecules for Ni-1 atom; while four O atoms from carboxyl groups of four TCPE ligands and two μ₂-O atoms for Ni-2 atom bridged to Ni-1 and Ni-1#5 (symmetry codes: #5, x, −y, −z + 3) forming a linear-shaped trinuclear Ni₃ cluster (Fig. 1b). The Ni–O bonds and O–Ni–O angles are in the ranges of 1.985(6)–2.071(19) Å and 85.1(6)–180.0(4)°, respectively. Each Ni₃ cluster is connected to four different deprotonated TCPE ligands and further extended to construct a three-dimensional framework (Fig. 1c). The four carboxylic groups within one ligand exhibit two types of coordination modes, in which two of them retain one free oxygen atom. When viewed along the b axis, Ni-1 exhibits a large rhombic channel with a height of 8.56 Å and width of 27.21 Å. In the absence of guest molecules, the effective free volume of Ni-1 was calculated using PLATON to be 50.9% of the crystal volume. The Ni₃ clusters are regarded as eight-point-connected nodes, and the ligands are considered to be the linkers. The overall framework can be regarded as a scu topology with the Schläfli symbol {4^16.6^12}{4^4.6^2}2 (Fig. 1d).

Fig. 1 (a) The asymmetric unit of Ni-1; (b) the linear trinuclear Ni₃ cluster in Ni-1; (c) perspective view of the three-dimensional framework along the b axis; (d) the schematic representation of Ni-1 network as tiling. Colour code: Ni, pink; O, yellow; C, gray. The H atoms and lattice solvents are omitted for clarity.

Structure of Ni-2. Ni-2 belongs to the trigonal space group P6/mmm. The asymmetric unit of Ni-2 contains a quarter of one Ni ion, one-eighth of a TCPE ligand, a quarter of a coordinated water molecule and one-twelfth of a μ₃-O atom (Fig. 2a). The Ni–O bonds and O–Ni–O angles are in the ranges of 1.985(6)–2.071(19) Å and 85.1(6)–180.0(4)°, respectively. Because of the different symmetry, Ni-2 contains a trinuclear Ni₃ cluster in which all the Ni atoms are coordinated to four different TCPE ligands, one coordinated H₂O molecule and one μ₃-O bridging atom. The different construction of the Ni₃ cluster may due to the different steric effects or polarity of solvent systems towards rotation of the carboxylic ligand. TCPE ligands serve as linkers to connect each Ni₃ cluster and all the eight oxygen atoms in the carbonyl groups are coordinatively saturated. Due to the partially twisted ethyl core and multiple rotational phenyl rings of the ligands, six Ni₃ clusters and six TCPE ligands are linked to form an enclosed ring with S₆ symmetry and further extended to three-dimensional frameworks without any interpenetration. The diameter of the 1D honeycomb pores is ca. 26 Å, along the c axis. Interestingly, there are several small pores on the side-wall of the frameworks with a diameter of ca. 9.0 Å (Fig. 3). The PLATON program analysis demonstrates that the pores constitute 71% of the crystal volume, which is much greater than Ni-1. If each Ni₃ cluster is considered as a six-connected node and each TCPE ligand as a four-connected planar linker, the overall frameworks can be regarded as a stp topology with the Schläfli symbol {4^4.6^2}3{4^9.6^6}2 (Fig. 2d). Notably, the 3D hexagonal-shaped framework of Ni-2 is similar to the configuration of Ni-TCPE1 that has been reported previously by our group, but exhibits a discrete single-walled nanotube.⁴⁰ We considered that the main difference of the assembly corresponds to the different metal cluster and the coordination fashion of carbonyl groups in TCPE ligands. The coordinated DMF molecules on the Ni₂ units vastly blocked the coordination of extra ligands to extend the architectures.

Fig. 2 (a) The asymmetric unit of Ni-2; (b) the trinuclear Ni₃ cluster in Ni-2; (c) perspective view of the three-dimensional framework along the c axis; (d) the schematic of Ni-2 network as tiling.

Fig. 3 Two types of cavities showed in Ni-2 with different sizes.

Solvent stabilities. To examine the solvent stabilities of Ni-1 and Ni-2, the samples were soaked in several common solvents, including water, DMF, C₂H₅OH, C₂H₂Cl₂, DMSO and 1,4-dioxane, for four days. All the PXRD patterns of the two complexes after being dispersed in these solvents were consistent with the simulated ones, indicating that the frameworks are preserved, as shown in Fig. 4. In addition, the IR spectra of the samples in different solvents exhibited the characteristic peaks of the fresh samples without significant variation (Fig. S7 and S8†). These results fully suggested that Ni-1 and Ni-2 can be highly stable in these solvents to maintain the skeletons of their framework structures. The high chemical stability in multiple solvents endows these MOF materials with ideal potential as heterogeneous catalysts for various valuable organic transformations.

Fig. 4 The family of PXRD patterns of fresh crystals Ni-1 (a), Ni-2 (b), their calculated patterns based on the single-crystal simulation and the samples after being dispersed in different solvents.

Gas adsorption studies. To evaluate the porosity of Ni-1 and Ni-2, gas adsorption experiments were carried out. Before the measurement, the crystal samples were immersed in ethanol to exchange the uncoordinated solvent molecules. Then, the outgassing process was conducted at 100 °C under high vacuum for 6 h to obtain the activated samples. The sorption isotherms of N₂ and CO₂ for both activated samples are shown in Fig. 5. N₂ adsorption isotherms at 77 K of both complexes indicate that they exhibit type I sorption behavior with a saturated uptake of 206.4 and 279.1 cm³ g⁻¹ at 1.0 bar, respectively, which corresponds to the typical microporous materials. The Brunauer–Emmett–Teller (BET) surfaces calculated from the adsorption isotherms are 573.8 m² g⁻¹ for Ni-1 and 813.2 m² g⁻¹ for Ni-2. Almost no N₂ adsorption was observed at 273 K for both complexes. The CO₂ adsorption measurements for the complexes at 273 and 298 K are also reversible and show a steady rise with the uptake of 62.1 cm³ g⁻¹ (corresponding 2.77 mmol g⁻¹) and 50.0 cm³ g⁻¹ (2.23 mmol g⁻¹) for Ni-1 at 1 bar, respectively. The CO₂ sorption uptake of Ni-2 at these two temperatures are similar to those of Ni-1 and display uptakes of 74.8 cm³ g⁻¹ (corresponding 3.34 mmol g⁻¹) and 49.6 (corresponding 2.21 mmol g⁻¹) cm³ g⁻¹ at 1 bar, respectively. These results not only suggest the porosity of our Ni-based MOF materials and possibility of adsorbing the gas molecules within their pores,^41,42 but they also confirm high CO₂ uptake ability, facilitating them to activate and catalyze CO₂ molecules as C1 source in the catalysis system.

Fig. 5 Gas adsorption isotherms of Ni-1 (a) and Ni-2 (b) of N₂ at 77 K, CO₂ measured at 273 and 298 K, respectively. Filled shape, adsorption; open shape, desorption.

Catalysis performance of cycloaddition with CO₂. Considering the high solvents stabilities and gas uptake ability discussed above, we believe that both the complexes Ni-1 and Ni-2 meet the basic prerequisites as ideal platforms for heterogeneous catalysis involving CO₂ transformations. After the activation treatments, the dispersive Ni ions perform as Lewis active sites. The large and accessible nanoscopic pores provide suitable space for the contact of carbon dioxygen and organic substrate molecules via adsorption and interaction. Thus, the cycloaddition reactions from CO₂ and epoxides were carried out under ambient conditions using Ni-1 and Ni-2 as heterogeneous catalysts. In a typical experiment, the reactions were conducted in an autoclave reactor using epoxide (10 mmol) with 1 atm pressure of CO₂ under a solvent-free environment at 313 K. In the presence of 0.5 mmol of tert-butylammonium bromide (TBABr), the loading of 0.3 mol% ratio of catalysts (based on the Ni sites) afforded almost complete conversions for the cycloaddition of CO₂ and propylene oxide to propylene carbonate after 40 h (Table 1, entry 1), with a yield of 98.4% for Ni-1 and 99.0% for Ni-2. Control experiments suggested that no detectable conversion occurred in the absence of Ni-catalysts or TBABr. An impressive decrease in the conversion was observed when the phenyl group in styrene oxide was substituted for a phenoxymethyl group, as illuminated by the 67.8% yield for Ni-1 and 69.5% for Ni-2 (Table 1, entry 2). The comparatively lower yield of styrene oxide is attributable to the low reactivity at its β-carbon center.⁴³ Other methoxy-substituted epoxide substrates could react with CO₂ to provide the desired products in good to excellent yields (Table 1, entry 3–5). However, on increasing the size of the epoxide (Table 1, entry 6), the bulky substrate 1f gave less than 5% yield under the same conditions. The catalytic activity was quite low and provides evidence that the catalysis occurs within the pores of the frameworks.^44–46 Although the size of the substrate 1f is slightly smaller than the size of the pores in Ni-1 and Ni-2, the size of the cycloaddition product formed inside the pores is larger and therefore, the diffusion to separate the product from the system was limited and reaction was reduced. Notably, the two catalysts contain the same amount of catalytic sites; however, the activity of Ni-1 is significantly lower than that of Ni-2, displaying size-selective catalytic performance. Recyclability is an essential feature of any catalyst considered for use in industrial applications. The recycled catalyst from the reaction was further employed for successive runs with freshly added co-catalyst. The yields of cyclic carbonates were not significantly affected in further cycles (Table S2†).

Table 1 Cycloaddition reactions catalysed by two catalysts from different substituted epoxides with CO₂^a

Entry	Substrate	Product	Yield^b (%)
			Ni-1	Ni-2
a Reaction conditions: epoxide (10 mmol), catalyst (30 μmol, based on the active Ni sites), and TBABr (0.5 mmol) under carbon dioxide (1 atm), 313 K and 40 h.b The yields were determined by ^1H NMR analysis using 1,3,5-trimethylbenzene as internal standard for integration.c 1f (1 mmol) with 2 mL CDCl3 as solvent.
1			98.4	99.0
2			67.8	69.5
3			96.9	97.8
4			88.9	89.6
5			83.7	85.5
6^c			—	4.9

---

A tentative mechanism based on these observations and previous reports for the cycloaddition of epoxide and CO₂ in the presence of Ni-MOFs co-catalyzed by TBABr is illustrated (Scheme S1†).^47–49 After the coordinating water molecules are removed, the Ni active sites in the pores might be activated and can serve as Lewis acid catalytic sites to activate the epoxy ring through the oxygen atom. Then, Br⁻ generated from TBABr attacks the less-hindered methylene carbon atom of the activated epoxide to open the epoxy ring. Concomitantly, the activated ring-opening intermediate reacts with CO₂ to form an alkylcarbonate anion. Finally, ring-closure through an intramolecular nucleophilic attack by the oxyanion at the carbon center of CO₂ occurs to generate a cyclic carbonate and regenerate the catalyst.

Conclusions

In conclusion, two types of Ni(II)-based coordination frameworks have been solvothermally synthesized via the elaborate regulation of solvent driven self-assembly, incorporating the tetraphenylethylene moiety as the backbone. These two complexes both contain a trinuclear metal cluster, but with different styles and construction modes, probably due to the different steric effects or polarity of the different solvent systems toward the rotation of the carboxylic ligand. IR and XRD spectra demonstrate the significant stabilities of the two frameworks toward various solvents. Importantly, both of them could act as effective heterogeneous catalysts for cycloaddition of CO₂ with epoxides forming cyclic carbonates under ambient conditions and provide appropriately sized pores to encapsulate the reaction substrates. The size-selectivity of the substrates suggested that the catalytic processes occurred within the pores of the catalysts. This study suggests a new approach for the solvent driven construction of different porous frameworks for use as efficient heterogeneous catalysts.

Acknowledgements

We gratefully acknowledge the financial support from the National Natural Science Foundation of China (21531001 and 21421005).

Notes and references

1. A. G. Slater and A. I. Cooper, Science, 2015, 348, aaa8075.
2. J. P. Zhang, Y. B. Zhang, J. B. Lin and X. M. Chen, Chem. Rev., 2012, 112, 1001–1033.
3. J. P. Sculley and H. C. Zhou, Angew. Chem., Int. Ed., 2012, 51, 12660–12661.
4. H. Furukawa, U. Müllerv and O. M. Yaghi, Angew. Chem., Int. Ed., 2015, 54, 3417–3430.
5. Y. B. He, W. Zhou, G. D. Qian and B. L. Chen, Chem. Soc. Rev., 2014, 43, 5657–5678.
6. J. W. Liu, L. F. Chen, H. Cui, J. Y. Zhang, L. Zhang and C. Y. Su, Chem. Soc. Rev., 2014, 43, 6011–6061.
7. M. Du, C. P. Li, C. S. Liu and S. M. Fang, Coord. Chem. Rev., 2013, 257, 1282–1305.
8. Z. Chang, D. H. Yang, J. Xu, T. L. Hu and X. H. Bu, Adv. Mater., 2015, 27, 5432–5441.
9. J. S. Qin, D. Y. Du, M. Li, X. Z. Lian, L. Z. Dong, M. Bosch, Z. M. Su, Q. Zhang, S. L. Li, Y. Q. Lan, S. Yuan and H. C. Zhou, J. Am. Chem. Soc., 2016, 138, 5299–5307.
10. J. Lü, J. X. Lin, M. N. Cao and R. Cao, Coord. Chem. Rev., 2013, 257, 1334–1356.
11. M. Li, D. Li, M. O. Keeffe and O. M. Yaghi, Chem. Rev., 2014, 114, 1343–1370.
12. K. Shen, M. D. Zhang and H. G. Zheng, CrystEngComm, 2015, 17, 981–991.
13. G. Sneddon, A. Greenaway and H. H. P. Yiu, Adv. Energy Mater., 2014, 4, 1301873.
14. Y. G. Zhang and D. S. W. Lim, ChemSusChem, 2015, 8, 2606–2608.
15. Q. Wang, J. F. Bai, Z. Y. Lu, Y. Pan and X. Z. You, Chem. Commun., 2016, 52, 443–452.
16. K. Huang, C. L. Sun and Z. J. Shi, Chem. Soc. Rev., 2011, 40, 2435–2452.
17. S. Pulla, C. M. Felton, P. Ramidi, Y. Gartia, N. Ali, U. B. Nasini and A. Ghosh, J. CO₂ Util., 2013, 2, 49–57.
18. P. Dürre and B. J. Eikmanns, Curr. Opin. Biotechnol., 2015, 35, 63–72.
19. J. Bayardon, J. Holz, B. Schäffner, V. Andrushko, S. Verevkin, A. Preetz and A. Börner, Angew. Chem., Int. Ed., 2007, 46, 5971–5974.
20. M. Aresta, A. Dibenedetto and A. Angelini, Chem. Rev., 2014, 114, 1709–1742.
21. P. Markewitz, W. Kuckshinrichs, W. Leitner, J. Linssen, P. Zapp, R. Bongartz, A. Schreiber and T. E. Muller, Energy Environ. Sci., 2012, 5, 7281–7305.
22. M. North and R. Pasquale, Angew. Chem., Int. Ed., 2009, 48, 2946–2948.
23. C. Martín, G. Fiorani and A. W. Kleij, ACS Catal., 2015, 5, 1353–1370.
24. X. B. Lu and D. J. Darensbourg, Chem. Soc. Rev., 2012, 41, 1462–1484.
25. Y. Xie, T. T. Wang, X. H. Liu, K. Zou and W. Q. Deng, Nat. Commun., 2013, 4, 1960.
26. W. Y. Gao, Y. Chen, Y. H. Niu, K. Williams, L. Cash, P. J. Perez, L. Wojtas, J. F. Cai, Y. S. Chen and S. Q. Ma, Angew. Chem., Int. Ed., 2014, 53, 2615–2619.
27. J. Zheng, M. Y. Wu, F. L. Jiang, W. P. Su and M. C. Hong, Chem. Sci., 2015, 6, 3466–3470.
28. R. Babu, A. C. Kathalikkattil, R. Roshan, J. Tharun, D. W. Kim and D. W. Park, Green Chem., 2016, 18, 232–242.
29. M. Yoon, R. Srirambalaji and K. Kim, Chem. Rev., 2012, 112, 1196–1231.
30. Y. Liu, W. M. Xuan and Y. Cui, Adv. Mater., 2010, 22, 4112–4135.
31. J. Y. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. T. Nguyen and J. T. Hupp, Chem. Soc. Rev., 2009, 38, 1450–1459.
32. J. Dong, P. Cui, P. F. Shi, P. Cheng and B. Zhao, J. Am. Chem. Soc., 2015, 137, 15988–15991.
33. H. Q. Xu, J. H. Hu, D. K. Wang, Z. H. Li, Q. Zhang, Y. Luo, S. H. Yu and H. L. Jiang, J. Am. Chem. Soc., 2015, 137, 13440–13443.
34. T. Zhang and W. B. Lin, Chem. Soc. Rev., 2014, 43, 5982–5993.
35. N. B. Shustova, B. D. McCarthy and M. Dincă, J. Am. Chem. Soc., 2011, 133, 20126–20129.
36. SMART Data collection software (version 5.629), Bruker AXS Inc., Madison, WI, 2003.
37. SAINT, Data reduction software (version 6.45), Bruker AXS Inc., Madison, WI, 2003.
38. G. M. Sheldrick, SHELXTL97, Program for Crystal Structure Solution, University of Göttingen, Göttingen, Germany, 1997.
39. A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13.
40. Z. Zhou, C. He, J. H. Xiu, L. Yang and C. Y. Duan, J. Am. Chem. Soc., 2015, 137, 15066–15069.
41. X. Q. Wang, L. L. Zhang, J. Yang, F. L. Liu, F. N. Dai, R. M. Wang and D. F. Sun, J. Mater. Chem. A, 2015, 3, 12777–12785.
42. S. N. Wang, T. T. Cao, H. Yan, Y. W. Li, J. Lu, R. R. Ma, D. C. Li, J. M. Dou and J. F. Bai, Inorg. Chem., 2016, 55, 5139–5151.
43. A. C. Kathalikkattil, D. W. Kim, J. Tharun, H. G. Soek, R. Roshan and D. W. Park, Green Chem., 2014, 16, 1607–1616.
44. P. Z. Li, X. J. Wang, J. Liu, J. S. Lim, R. Q. Zou and Y. L. Zhao, J. Am. Chem. Soc., 2016, 138, 2142–2145.
45. W. Y. Gao, C. Y. Tsai, L. Wojtas, T. Thiounn, C. C. Lin and S. Q Ma, Inorg. Chem., 2016, 55, 7291–7294.
46. Z. H. Xu, L. L. Han, G. L. Zhuang, J. Bai and D. Sun, Inorg. Chem., 2015, 54, 4737–4743.
47. L. Liu, S. M. Wang, Z. B. Han, M. L. Ding, D. Q. Yuan and H. L. Jiang, Inorg. Chem., 2016, 55, 3558–3565.
48. Y. H. Han, Z. Y. Zhou, C. B. Tian and S. W. Du, Green Chem., 2016, 18, 4086–4091.
49. M. H. Beyzavi, R. C. Klet, S. Tussupbayev, J. Borycz, N. A. Ver-meulen, C. J. Cramer, J. F. Stoddart, J. T. Hupp and O. K. Farha, J. Am. Chem. Soc., 2014, 136, 15861–15864.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental and catalysis details, characterization data for the new compounds. CCDC 1502000 and 1502001. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6ra22971a

---