Metal–organic framework catalysis

The chemical mutability, bespoke pore structures and large, readily accessible internal surface areas of metal–organic frameworks (MOFs) facilitate possible applications in heterogeneous catalysis.¹ As a demonstration of the suitability of MOFs in this area, heterogeneous catalysis was postulated in Robson's seminal work on diamondoid coordination polymers 27 years ago² and realised by Fujita using a 2D square-based network material shortly afterwards.³ Realisation of industrially-relevant heterogeneous catalysis by MOFs would be of momentous significance as globally the catalyst market is of the order of 20 billion USD annually.⁴ This in turn leverages the production commodity chemicals, fuels and products worth at least a thousandfold this value.⁴

In addition to the underlying building blocks of MOFs and the frameworks themselves possessing sites of intrinsic catalytic potential (i.e. Lewis acidic metal sites, organocatalytic moieties, pore size and shape),¹ the framework can also undergo post-synthetic modification or metallation (PSM, PSMet) to incorporate catalytically competent chemical functionality.^5–7 Notable among these methods are strategies to incorporate various organocatalytic units or to structurally replicate molecular homogenous catalysts, particularly those closely related to industrially used examples. Utilising the MOF as a support for catalytically active nanoparticles or even site isolated molecular species is an extension of this approach.

The capacity to carefully manipulate the size and shape of the pore space available within a MOF network further engenders them toward ideas of using the pore network to discriminate reagents or direct reactivity. Furthermore, the tailoring of the pore network can be used to facilitate excellent mass transport and this can be further enhanced by structuralising the MOF onto a hierarchically porous support.⁸ This vast array of potential different catalytic moieties within MOFs, often with several available in a single material, facilitates the development of synergistic catalysts and materials capable of catalysing tandem reactions, a subject that has recently been reviewed.⁹ High surface areas in turn offer a high density of active sites and potentially a more effective catalyst. Much activity, due in part to restricted thermal stability with respect to other competitor materials, like zeolites, and a lack of permanent porosity for some delicately functionalised examples, has meant considerable attention has been placed on catalysis of condensed-phase reactions. However, as thermal and chemical stability have been improved, the full potential of the microporosity of MOFs has been realised for gas phase reactions.¹⁰

A short summary of each paper submitted to the metal–organic framework catalysis issue is presented below. We would like to take this opportunity to thank all of the authors for their contributions, which admirably cover the scope of MOF catalysis and point to further opportunities and aspects that must be addressed before MOF-based materials can be realistically used as catalysts.

Due to protective environment within the pore, coordinatively unsaturated metal sites can often be unveiled and utilised in MOFs wherein the equivalent molecular species would undergo dimerisation or decomposition. Such unsaturated metal sites were used by Fujita for cyanosilyliation reactions in the first report of MOF heterogeneous catalysis. In their Highlight in this issue, Hu and Zhao review the recent development in the design and synthesis of MOFs with Lewis acid sites, and the application of these moieties in heterogeneous catalysis, including in addition, redox and hydrogenation reactions. Several strategies to form Lewis acidic sites are reviewed with particular attention given to the characterisation techniques used to probe these moieties.

Examples of Lewis acidic sites in MOF catalysis are described by the groups of Fischer and Ibarra in this issue. Fischer et al. reported cyanosilyliation of benzaldehyde catalysed by Lewis acid sites in defect engineered UiO-66. Acetic acid and trifluoroacetic acid were used to modulate UiO-66 synthesis and install an increasing number of defects. The chemistry of these sites was probed by water adsorption measurements allowing a direct measure of their hydrophilicity, which in turn correlates well with their catalytic performance. Notably the defect engineered samples were better catalysts that the parent (defect-free) MOF, although it was uncertain whether this was specifically due to the greater density of Lewis acid sites or the improved diffusion.

In their Communication in this issue, Sánchez-González et al. utilise the Lewis acid sites on the periphery of a Cu(II) paddlewheel based metal–organic polyhedron (MOP) to oxidise trans-ferulic acid to vanillin. Analogous chemistry has been reported with comparable Lewis acidic sites in MOFs like HKUST-1 (Cu₃(BTC)₂). Optimisation of the conditions leads to good conversions with the reaction being selective for the formation of vanillin and, despite the relatively weak bonding within the crystal lattice, the MOP can be reused (albeit with some loss of conversion over 5 cycles). While perhaps not the catalyst of choice for industrial use, such MOP-based catalysts might facilitate further mechanistic studies that cannot be readily undertaken on MOFs.

As noted, coordinatively unsaturated sites within MOFs are common sites of catalytic activity. A number of MOFs possess open metal sites, including HKUST-1 mentioned above and the two MOFs that are the subject of the Communication by Ameloot and co-workers, namely UTSA-74 and MOF-74(Zn). For such MOFs, stability to reaction conditions, which may include solvents often used in MOF synthesis or water, is required for applications in catalysis. The authors show that the recently identified UTSA-74, [Zn₂(dobdc)] (dobdc = 2,5-dioxidobenzene-1,4-dicarboxylate), transforms to its well studied polymorph MOF-74(Zn) upon exposure to water as a liquid or vapour. These results suggest that experiments to utilise the two coordinatively unsaturated sites in UTSA-74 for catalysis will require very careful control of moisture, which is not viable for many reactions.

Taking advantage of the Brønsted basicity of amine-functionalised linkers, the Al-MOFs CAU-1-NH₂ and CAU-10-NH₂ were studied as catalysts for the Knoevenagel condensation of benzaldehyde and malononitrile. Mixed linker CAU-1-NH₂/H and CAU-10-H MOFs were also studied. As a consequence of their differing topologies the two families of materials possess different pore sizes and shapes and hence display different selectivities in the formation of the desired α,β-unsaturated compound and avoid further conversion (also the Lewis acidic sites within the MOF can catalyse the formation of by-products). CAU-1-NH₂ was identified as the best catalyst due to its selective formation of benzylidene malononitrile; CAU-1-NH₂/H, CAU-10-NH₂ and CAU-10 all lead to good conversions of benzaldehyde but facilitate the formation of benzaldehyde diethyl acetal (as the exclusive product in the case of CAU-1-NH₂/H due to the greater availability of Lewis acid sites). The authors attribute the reactivity profile to the accessibility of amine sites in CAU-1-NH₂.

Demonstrating an extension of the ideas for developing MOF heterogeneous catalysis, the groups of Gascon and Llabrés i Xamena report the heterogenisation of a molecular catalyst for transfer hydrogenation within a porous organic framework (POF). A triazine based POF was formed from 2,6-pyridinedicarbonitrile and converted to an Ir@covalent triazine framework (Ir@CTF) by treatment with [IrCp*Cl₂]₂. Using this system, which can be idealised as a [IrCp*(solvent)] moiety coordinated to bipyridine moieties within the CTF, isomerisation of allylic alcohols to saturated ketones was examined. The catalyst is air and moisture stable and able to achieve comparable conversion and improved turnovers compared to related Ir(III) salts or a UiO-68 bound Iridium N-heterocyclic carbene catalyst. The catalyst was shown to be recyclable although the selectivity for the saturated ketone product versus the saturated alcohol (transfer hydrogenation from the isopropanol solvent) decreased at longer reaction times.

Supporting catalytically active species inside the pore networks of MOFs has been a longstanding focus. In two Full Papers Xie et al. and the group of Balula and Cunha-Silva report catalysis by Pd nanoparticles and the iron-substituted polyoxometalate (POM) TBA₄[PW₁₁Fe(H₂O)O₃₉] (PW₁₁Fe), respectively. In their Article Xie et al. report the formation of Pd nanoparticle loaded MOF Pd@Zn₂(azoBDC)₂(dabco). The heterogeneous catalyst was formed by linker exchange and pore expansion of a related pillared MOF Zn₂(BDC)₂(dabco), Pd impregnation and ethylene glycol reduction to give the Pd@Zn₂(azoBDC)₂(dabco) catalyst. The authors examined the reduction of nitroaromatics to amines using NaBH₄ as a hydride source, specifically looking at the case of 4-nitrophenyl. The catalysts showed good conversion, excellent recyclability and no detectable leaching of Pd.

In a related approach using a trapped species, Balula and Cunha-Silva et al. reported encapsulation of the POM PW₁₁Fe into the MOF NH₂-MIL-101(Fe) to yield PW₁₁Fe@NH₂-MIL-101(Fe). The POM is a previously studied catalyst for the ring-opening of styrene oxide with amines, but suffers from issues of recovery and reuse. The authors investigated the activity of the pure MOF NH₂-MIL-101(Fe) revealing low conversions in this reaction after 1 hour (the reaction goes to completion at long reaction times >24 hours), where the catalysis presumably occurs at coordinatively unsaturated defect sites in the MOF. POM-impregnated PW₁₁Fe@NH₂-MIL-101(Fe) catalyses this reaction to completion within an hour, is reusable and stable.

In a series of Highlights authors in the special issue examine biomass upgrading, photocatalysis, electrocatalysis and biocatalysis. Herbst and Janiak critically compare the potential of MOFs and other catalysts (e.g. zeolites) for biomass transformation and conversion (valorisation) to chemical feedstocks. The authors note that catalysis by MOFs is investigated to a lesser extent than other competitor materials (shown as a share of publications on these materials), with biomass conversion only being a relatively recent focus for MOFs. This can in part be attributed to the stability of MOF catalysts to water, which is an essential part of the often acid catalysed hydrolysis reactions occuring in biomass conversion. They find that in some instances existing MOF catalysts can favourably compete with other potential catalysts (e.g. in levulinic acid conversion to ethyl levulinate), but in other instances they are outperformed (e.g. glucose dehydration and cellulose hydrolysis). Product specificity can also be engendered by certain MOF catalysts due to their capacity for structural tuning.

Two Highlights on photocatalysis are included in the issue. Cohen and co-workers examine how MOFs have attracted increasing attention for applications in heterogeneous photocatalysis of organic transformations. They report how the SBUs can be used as sites of photocatalytic reactions, in particular promising work with Fe-based MOFs; how the linkers can facilitate incorporation of photocatalytic moieties (e.g. tris(2,2′-bipyridine)ruthenium(II) or porphyrins); and how photocatalytic guests can be utilised. Additionally, the review examines MOF composites, integrating noble metal nanoparticles or traditional photoactive semiconductors with MOFs, to show the advantageous and synergistic properties of the combined materials. Gascon and van der Meen's Highlight advocates greater fundamental characterisation of the MOFs used in photocatalytic solar fuel production as a means to enhance efficiency. They examine the MOFs thus far investigated for photocatalysis paying particular attention to the photophysics of d⁰ MOFs, which are very often based on Ti⁴⁺ and Zr⁴⁺ containing nodes. Alternate “ship-in-a-bottle” type examples are also considered whereby a photocatalyst is impregnated into the pores as a guest. In spite of the challenging results for photocatalytic performance the authors note a number of the recently prepared, conductive MOFs have yet to be tested for photocatalysis.

In their Highlight Solomon et al. make the case for MOFs as electrocatalysts to convert CO₂ and H₂O to fuels. Their contention is that electrocatalytic transformations, which can occur in Faradaic efficiencies of up to 100%, can be coupled with renewable energy sources and that heterogeneous electrocatalysts deposited on electrodes offer a number of chemical and practical benefits over homogenous electrocatalysts. The Highlight examines MOFs displaying intrinsic electrocatalytic properties with a focus on the CO₂ reduction reaction and water splitting (oxygen evolution, hydrogen evolution reactions). The conclusions are that MOFs provide structural features that make them excellent electrocatalysts but the inherent stability of the frameworks and product selectivity are ongoing challenges.

One of the earliest strategies to engender catalytic potential to MOFs was to use them as a porous framework to support and stabilise catalytically competent guests. Recent work on enzyme@MOF composites similarly seeks to stabilise and protect enzymes within a porous structure and to take advantage of these highly selective and efficient catalysts. Such protective conditions would enable these inherently delicate structures to be used under biologically challenging conditions, i.e. in the presence of denaturants, at high temperatures. The Highlight by Farha focuses on the synthetic methods to prepare enzyme@MOF composites and their subsequent catalytic applications. So-called de novo encapsulating methods (co-precipitation and biomimetic mineralisation)^8b,11 are examined, as are methods to post-synthetically encapsulate enzymes into large pore MOFs. Both methods are shown to have their advantages and disadvantages but as yet reliable, free enzyme-like catalysis has not been reliably observed. Nonetheless, significant potential exists for such biocomposites as catalysts.

A longstanding principle of MOF chemistry is the ability to control the density of active sites by incorporating differing levels of interpenetration into a material. Two closely related studies have been reported by the groups of Yanli Zhao and Ma to examine MOFs possessing either different degrees of interpenetration or different topologies, respectively, and their effects on CO₂ conversions. In the work of Zhao, two MOFs with 2-fold or 4-fold interpenetrated structures were studied. Both materials have unsaturated Cu Lewis acid centres capable of catalytic conversion of CO₂ but, due to the differing levels of interpenetration, quite different pore sizes and thereby catalytic behaviour. The less dense 2-fold interpenetrated structures has an accessible pore network, much higher CO₂ uptake and notably better conversion of CO₂ into cyclic carbonates than the 4-fold interpenetrated material. The data clearly indicates that catalysis in the 4-fold interpenetrated material occurs predominantly on the crystal surface whereas the 2-fold interpenetrated facilitates catalysis within the pores. With similar intent Ma et al. manipulated the conditions of a MOF synthesis to form either the lvt or nbo (as in the NOTT series) topologies. Both networks provide access to coordinatively unsaturated Cu(II) Lewis acid centres but again with different pore sizes, N₂ uptakes and BET surface areas. The yields for the reaction of CO₂ with epoxides to form the corresponding cyclic carbonates were then assessed, with the nbo topology showing marginally higher yields. The differences between the yields of cyclic carbonate for the lvt and nbo topologies are less pronounced than in the previous work by Zhao presumably due to both topologies still being capable of catalysis in the pores (the 4-fold interpenetrated material is essentially non porous).

To provide greater accessibility of catalytic sites and to improve reactant/product mass transport, Coskun and co-workers report CO₂ conversion by a hierarchically porous form of ZIF-8 attained by conversion of nanostructured ZnO. In their approach ZnO nanoparticles were self-assembled with polymer beads to polymer@ZnO composites, with the polymer template then removed by pyrolysis. The resulting ZnO nanostructures were then reacted with 2-methylimidazole to form a range of ZiF-8/ZnO nanostructures with differing hierarchically porous structures. The effect of the structuring on reactivity was assessed by looking at the catalytic activity of ZIF-8/ZnO in the conversion of CO₂ using epichlorohydrin to form chloropropene carbonate. ZIF-8 structured on a macroporous nanostructure had better conversion than the mesoporous material and both were an improvement over ZIF-8 in terms of product selectivity and reusability (albeit to a limited extent in the latter).

In the field of MOF catalysis indirect evidence is typically provided as to the type and location of the catalytic centres (guest reactant adsorption data, pore size measurements, crystal structure data etc.). In their Communication Kubarev and Roeffaers use super-resolution fluorescence microscopy to map surface acid–base catalytic activity for ZIF-8. In their experiments the catalytically active regions of MOF crystals can be visualised by conversions of non-fluorescent reactants into fluorescent products. Remarkably individual turnovers can be resolved with resolution in the tens of nm range. Using this approach the authors were able to detect catalysis only on the outer crystal surface and at defects, in support of earlier studies on catalysis by ZIF-8.

Surface coatings on MOF crystals and conversely MOF coatings on other compounds provides access to composite materials with enhanced properties or new opportunities for applications. In their Communication Kishimoto et al. use a MOF, iron(II) oxalate (Fe(ox)) to modify the surface of LiCoO₂, a catalyst for the oxygen evolution reaction (OER). The process involves step-by-step exposure of LiCoO₂ particles to solutions of oxalic acid and an Fe(II) salt. The nanolayer coating quickly reaches a maximum thickness after a few cycles according to energy-dispersive X-ray spectroscopy data. While the catalytic properties of coated LiCoO₂ for oxygen evolution and oxygen reduction reactions were not assessed, the work shows that modification of inorganic nanoparticulate catalysts with nanolayers of MOFs can act to stabilise the material without changes in resistivity.

In an Article the groups of Matsuoka and Falcaro prepare magnetically recyclable catalysts by converting a Fe₃O₄ nanoparticle impregnated Cu-based ceramic into HKUST-1. The Fe₃O₄@HKUST-1 is formed in a room temperature conversion reaction and easily purified from HKUST-1 by-products by magnetic separation. Furthermore, the versatility of the synthetic method is shown by experiments to incorporate Pd nanoparticles during the conversion of the Cu-ceramic (giving Pd/Fe₃O₄@HKUST-1). The Lewis and Brønsted acidic sites (unsaturated Cu(II) sites and free carboxylate groups) of Fe₃O₄@HKUST-1 suggest applicability for a one-pot deacetalization–Knoevenagel condensation reaction. In these reactions the Fe₃O₄@HKUST-1 composite achieves good conversions and high yields of the benzylidene-malononitrile product. Furthermore the catalyst is reusable and easily separated from the reaction with a magnet. The Pd/Fe₃O₄@HKUST-1 composite was used for hydrogenation of 1-octene and again gave good results with facile separation.

Takahashi and coworkers report in an Article electrochemical sensing and catalysis by a Cu₃(BTC)₂ (HKUST-1) coating structuralised on gold electrodes. The Cu₃(BTC)₂ coating was formed by exposing an ethanolic solution of the linker to Cu(OH)₂ nanobelts which have been deposited on the gold electrodes. The conversion conditions can be carefully controlled to dictate the extent of conversion and the Cu₃(BTC)₂ particle size. The primary focus of the work was to demonstrate that sensors for size-selectively detecting electroactive analytes in the solution could be prepared (small iron complexes could diffuse through the Cu₃(BTC)₂ coating whereas sterically demanding complexes could not). Under basic conditions the MOF coating was tested for the electrocatalytic oxidation of glucose and showed a concentration sensitive chronoamperometric response.

Recognising that most MOFs are produced as powders yet applications, including those in catalysis, will require a material of a specific form, Dhainaut et al. undertook a study of the effect of MOF densification (as tablets) on the porosity and mechanical strength. This study was conducted for four materials (three from the UiO series and HKUST-1), including those that have explicitly featured in papers by other authors in this issue. The authors use mild compression to pelletise the MOFs (without binder), carefully and systematically identifying conditions (e.g. compression ramp speed, dwell time) to optimise the mechanical stability without loss of intrinsic properties, like porosity. Importantly, given the wide interest in HKUST-1, the authors identify conditions to densify this material.

In summary, it is clear that MOF catalysis is an active field as researchers continue to explore how the intrinsic properties of MOFs can be applied to long-standing challenges in catalysis. The collective work presented in this themed issue demonstrates that MOFs are an excellent platform for studying catalytic processes and offer rich opportunities for the development of novel catalytic systems.

References

1. H. Furukawa, K. E. Cordova, M. O'Keeffe and O. M. Yaghi, Science, 2013, 341, 1230444.
2. B. F. Hoskins and R. Robson, J. Am. Chem. Soc., 1990, 112, 1546–1554.
3. M. Fujita, Y. J. Kwon, S. Washizu and K. Ogura, J. Am. Chem. Soc., 1994, 116, 1151–1152.
4. J. A. Schaidle, S. E. Habas, F. G. Baddour, C. A. Farberow, D. A. Ruddy, J. E. Hensley, R. L. Brutchey, N. Malmstadt and H. Robota, in Catalyst, ed. J. Spivey and Y.-F. Han, Royal Society of Chemistry, Cambridge, 2017, vol. 29, pp. 213–281.
5. S. M. Cohen, Chem. Rev., 2012, 112, 970–1000.
6. J. D. Evans, C. J. Sumby and C. J. Doonan, Chem. Soc. Rev., 2014, 43, 5933–5951.
7. T. Islamoglu, S. Goswami, Z. Li, A. J. Howarth, O. K. Farha and J. T. Hupp, Acc. Chem. Res., 2017, 50, 805–813.
8. (a) S. Furukawa, J. Reboul, S. Diring, K. Sumida and S. Kitagawa, Chem. Soc. Rev., 2014, 43, 5700–5734; (b) R. Ricco, C. Pfeiffer, K. Sumida, C. J. Sumby, P. Falcaro, S. Furukawa, N. R. Champness and C. J. Doonan, CrystEngComm, 2016, 18, 6532–6542.
9. Y.-B. Huang, J. Liang, X.-S. Wang and R. Cao, Chem. Soc. Rev., 2017, 46, 126–157.
10. (a) See, for examples: Z. Li, N. M. Schweitzer, A. B. League, V. Bernales, A. W. Peters, A. Getsoian, T. C. Wang, J. T. Miller, A. Vjunov, J. L. Fulton, J. A. Lercher, C. J. Cramer, L. Gagliardi, J. T. Hupp and O. K. Farha, J. Am. Chem. Soc., 2016, 138, 1977–1982; (b) R. J. Comito, K. J. Fritzsching, B. J. Sundell, K. Schmidt-Rohr and M. Dincă, J. Am. Chem. Soc., 2016, 138, 10232–10237; (c) A. Burgun, C. J. Coghlan, D. M. Huang, W. Chen, S. Horike, S. Kitagawa, J. F. Alvino, G. F. Metha, C. J. Sumby and C. J. Doonan, Angew. Chem., Int. Ed., 2017 DOI:10.1002/ange.201611254; (d) E. D. Metzger, R. J. Comito, C. H. Hendon and Mircea Dincă, J. Am. Chem. Soc., 2017, 139, 757–762.
11. C. Doonan, R. Riccò, K. Liang, D. Bradshaw and P. Falcaro, Acc. Chem. Res., 2017, 50, 1423–1432.

---