Porous crystalline frameworks for thermocatalytic CO₂ reduction: an emerging paradigm

Abstract

Heterogeneous catalysts for CO₂ reduction derived from porous, crystalline frameworks have emerged as efficient systems with comparable activity and superior selectivity to their inorganic counterparts. The spatial arrangement of active sites in such catalytically active frameworks is critical to their performance in CO₂ reduction. This review presents a comprehensive and critical analysis of (thermal) CO₂ reduction over catalysts derived from porous, crystalline frameworks, whose structural and chemical diversity offers unprecedented opportunities to regulate reactivity. Thermodyamic considerations and the impact of process parameters on reaction intermediates, governing mechanisms for CO₂ reduction and catalyst stability are discussed. Strategies for leveraging the flexibility of porous, crystalline frameworks to improve their stability and promote CO₂ reduction are presented which include: use as sacrificial precursors to an active phase; integration within composites; and as hosts for nanoparticle encapsulation. Finally, future challenges and research prospects are highlighted.

---

Sunil Mehla

Sunil Mehla received his MSc in Chemistry from the Indian Institute of Technology Kanpur (2009) and MTech in Catalysis Technology, Chemical Engineering from the Indian Institute of Technology Madras (2012). He worked at HP Green R&D Centre of Hindustan Petroleum Corporation Limited for roughly 6 years (2018), researching and developing catalysts for isomerisation, cracking and diesel desulfurisation. He is currently pursuing his PhD at the Royal Melbourne Institute of Technology, Melbourne, under the supervision of Prof. Suresh Bhargava in the field of ordered noble metal nanoarrays for plasmonic sensing.

Ahmad E. Kandjani

Dr Ahmad Kandjani received his PhD in Applied Chemistry from RMIT in 2014. Since then, he has worked as a Research Fellow with the Nanotechnology and Sensing group. In 2017, he was awarded the prestigious Vice-Chancellor's Fellowship at RMIT University. Dr Kandjani has a strong track record with over 100 publications (h-index 23), including 77 journal articles, 1 book chapter and 2 patents (1 PCT patent). The impact of his research has been highlighted on the front cover of nine prestigious journals 9 times. His research interests include novel semiconductor materials, long ranged colloidal crystal, nano-patterning, and optoelectronic devices and sensors.

Ravichandar Babarao

Dr Ravichandar Babarao is a Senior Lecturer at RMIT University and a visiting scientist at CSIRO. He received his PhD from National University of Singapore with the Best PhD Thesis Award in 2010 and the Prosper.net Scopus Young Scientist Award in 2015. He has also received the Alexander von Humboldt visiting scientist fellowship, 2016 RACI Rennie Memorial Medal, Victoria Fellowship in Physical Science and the 2019 Peter Schwerdtfeger Award. With >60 publications, his research interests are the chemistry of novel nanoporous materials for energy, environmental and biomedical applications including storage, separation and drug delivery.

Adam F. Lee

Professor Adam Lee is a Professor of Sustainable Chemistry at RMIT University and previously held Chair appointments at Cardiff, Warwick, Monash, and Aston universities, and a prestigious EPSRC Leadership Fellowship. His research addresses the rational design of nanoengineered materials for clean catalytic technologies, and the development of in situ/operando methods providing molecular insight into surface reactions. He has authored >260 articles (h-index 65, 14 722 cites), received the 2011 McBain Medal of the Royal Society of Chemistry and SCI, and 2012 Beilby Medal and Prize of the Royal Society of Chemistry, IOM3, and SCI, and is Editor-in-Chief of Materials Today Chemistry.

Selvakannan Periasamy

Dr Selvakannan Periasamy has been working as a Research Fellow at RMIT University, since 2009. His major fields of interest are CO2 utilisation, methane activation, endothermic fuels for high speed flight vehicles, Surface Enhanced Raman Scattering and additive manufacturing in catalysis. He has 65 publications in international peer reviewed journals and a h-index of 33. Before joining RMIT, he worked as a Research Fellow at University of Paris-Sud, France (2007–2009) and as a Research Scientist at the Innovation Centre, Tata Chemicals Ltd, India (2005–2007). He received his PhD (2005) from the National Chemical Laboratory (University of Pune), India. He graduated from the American College, India with master's and bachelor's degrees in Chemistry (1996–2001).

Karen Wilson

Professor Karen Wilson is Professor of Catalysis at RMIT University and previously Chair of Catalysis and Research Director at the European Bioenergy Research Institute, Aston University, where she also held a prestigious Royal Society Industry Fellowship. Her research focuses on the development and scale-up of tunable, porous heterogeneous catalysts for green and sustainable chemistry and the utilisation of renewable resources in chemical processes. She has authored >260 journal papers (h-index 66, 14 819) and 20 book chapters, is Associate Editor of Sustainable Energy & Fuels, and serves on numerous Editorial Boards including Catalysis Today, Current Opinion in Green & Sustainable Chemistry and Energy & Environmental Materials.

Seeram Ramakrishna

Professor Seeram Ramakrishna, FREng is the Chair of Circular Economy Taskforce at the National University of Singapore (NUS). He is an editorial board member of Nature Scientific Reports. He is an elected Fellow of major professional societies and academies in Singapore, UK, India and USA. He is named among the World's Most Influential Minds and the Top 1% Highly Cited Researchers in Materials Science by Thomson Reuters and Clarivate Analytics. He has co-authored 1000 journal papers and 8 authored books received ∼110 000 citations and a h-index of 158. His research interests include innovations in sustainable materials and evaluation of circularity by life cycle assessment.

Suresh K. Bhargava

Distinguished Professor Suresh Bhargava is a world-renowned interdisciplinary scientist at RMIT University who is recognised for delivering research excellence that underpins significant industrial applications. His journal articles that number over 520 have been cited over 16 210 times. Out of his seven patents, five have gone to industries or licensed to commercialisation. Quoted as being among the top 1% world scientists in the resource sector, Professor Bhargava provides consultancy and advisory services to many government and industrial bodies around the world including BHP Billiton, Alcoa World Alumina, Rio Tinto and Mobil Exxon. He has been recognised through many prestigious national and international awards.

---

Broader context

The United Nations sustainable development goals have rallied environmental and energy scientists and engineers in pursuit of new technologies to initiate climate action and promote responsible consumption and production. Carbon dioxide (CO₂) emissions are a major contributor in climate change, however CO₂ capture and utilisation as a chemical feedstock could underpin a circular carbon economy. The transition from environmental pariah to abundant valuable feedstock requires new catalysts that can activate and selectively transform CO₂ into chemicals and fuels, through atom and energy economical processes. This review explores the emergence of porous crystalline frameworks as heterogeneous catalysts for the thermochemical reduction of CO₂ to small molecules. Elegant synthetic strategies enable structural diversity in porous crystalline frameworks, which in turn afford great control over the local chemical environment and reaction pathways available to CO₂, although challenges remain in operational stability and product selectivity. Scale-up of such catalytic technologies, and their application in innovative reactors and integration with renewable hydrogen sources, will likely drive significant growth in commercial thermocatalytic CO₂ reduction by 2030.

1.0 Introduction and background

The effects of rising anthropogenic CO₂ emissions on climate change are well documented,^1–3 and represent an imminent and existential threat to the biosphere requiring urgent and multipronged remediation strategies. The 21st Conference of the Parties (COP-21) to the United Nations Framework Convention on Climate Change, Paris 2015, tried to establish frameworks for national accountability and critical international comparisons to assist in meeting agreed targets, notably halving current CO₂ emissions by 2050 and restricting the increase in global average air temperature to <2 °C.^4,5 In addition, committees such as the Intergovernmental Panel on Climate Change, the Global Climate Change Initiative, the European Strategic Energy Technology Plan and the European Technology Platform for Zero Emission Fossil Fuel Power Plants have issued guidelines to direct progress in this field. Consequently, research into CO₂ mitigation has rapidly expanded, and now bridges physical and social sciences with engineering.

CO₂ removal from the atmosphere can be realised either using Carbon Capture and Sequestration (CCS), aimed at long-term geological storage or Carbon Capture and Utilisation (CCU) aimed at using CO₂ as a feedstock for production of substances with higher value. CO₂ Capture and Utilisation (CCU) has increasingly gained attention as a global warming remediation strategy due to potential applications in multiple sectors and the economic offset of investment costs by revenue from the sale of CO₂ derived products. CCU is reportedly 20–40 times more efficient than CO₂ Capture and Sequestration (CCS) over a 20 year timeframe.⁶ CCU can be applied to: (i) chemicals production; (ii) mineralisation; (iii) technological uses; (iv) enhanced biological use; and (v) fuels production.⁷ The first application is particularly appealing for the sustainable synthesis of commodity chemicals such as urea,^8,9 salicylic acid,¹⁰ inorganic carbonates,¹¹ formic acid,¹² acetic acid,^13,14 aromatic carboxylic acids,¹⁵ organic acyclic carbonates,^16,17 cyclic carbonates,^18,19 polycarbonates,^20–22 carbamates,^23–25 polyurethanes,²⁶ acrylates,²⁷ norbornane,²⁸ lactones^29,30 and pyrones,^31,32 currently produced from fossil resources. Note that production of such (comparatively) low volume organic chemicals cannot significantly mitigate CO₂ emissions,^33,34 and hence focus is shifting towards the synthesis of high volume/high demand fuels and chemicals such as oxygenates (alcohols and ethers),^35–37 alkanes,³⁸ olefins^39,40 and aromatics.^41,42 A recent study⁴³ identified 9 chemical products wherein a complete life cycle assessment suggested the use of CO₂ as a feedstock could reduce the impact of short- and long-term global warming. These include CO from dry reforming of methane, CO from reverse water gas shift reaction, methanol, methane, formic acid by hydrogenation, ethanol, dimethyl ether, dimethyl carbonate, and polyols. In contrast, production of formic acid electrocatalytically, dimethoxymethane or Fischer–Tropsch (FT) products from CO₂ was not beneficial for global warming, notably due to the high thermal energy demand for distillation of dilute formic acid solution (4.5 wt%) and associated cost of electricity and heat generation.⁴³

Use of CO₂ as a sustainable feedstock for chemicals and fuels is especially attractive as a process that can incorporate H₂ from renewable energy. The integration of such ‘renewable H₂’ permits the efficient synthesis of liquid transportation fuels with high combustion efficiencies and low pollutant emissions such as methanol, dimethyl ether, oxymethylene ethers and FT products.⁴⁴ When obtained from CO₂ reduction, CO and CH₃OH are valuable building blocks to other fuels and chemicals by established FT, Methanol to Hydrocarbons (MTH) and Methanol to Olefins (MTO) technologies. CO₂ conversion is energy intensive and often results in low product yields, hence initial research efforts focused on catalysts for CO and CH₃OH synthesis as the most accessible products based on technology readiness level and conversion metrics.⁴⁵ Subsequently, the direct reduction of CO₂ to methane (termed methanation) has received significant attention, with high activity catalysts offering ∼100% selectivity at modest reaction temperatures (∼250 °C). This mirrors the emergence of abundant, low cost renewable H₂, which offers an economically viable non-fossil route to hydrocarbons through the power-to-gas-concept; approximately 95 commercial power-to-gas projects are currently operational, generating ∼39 MW electricity worldwide.⁴⁶ Nevertheless, CO₂ methanation consumes large volumes of H₂ and the generated methane has a low volumetric energy density hindering transport applications wherein on-board storage issues remain. Many countries also have abundant reserves of natural gas, and for these the emphasis is on the development of methane conversion (rather than production) technologies⁴⁷ such as methane to ethane/ethylene, aromatics, methanol, or even methanesulfonic acid as a sustainable acidic solvent,⁴⁸i.e. chemical rather than fuel production. It is important to recall that methane is also a potent greenhouse gas, with ∼5 to 10 times warming potential of CO₂,⁴⁹ emitted in large quantities from natural sources such as forests, bushlands, swamps and hence any anthropogenic emissions must be tightly controlled. In any of the above scenarios, access to low cost, concentrated CO₂ and H₂ is critical, and the need to capture the former, a subject of intensive research at different scales.^50–52 Currently, the cost of renewable H₂ obtained from water electrolysis ($2.91–10.92 per kg)⁵³ still lags behind the cost of non-renewable H₂ ($1.35–3.5 per kg),⁵⁴ typically obtained from steam reforming of naphtha or natural gas. However, a recent report by the Hydrogen Council predicts the price of renewable H₂ will significantly drop ($1–3 per kg) by 2030 due to economies of scales in its production and distribution, and advances in electrolysis technology and the falling costs of electrolysers.⁵⁵ This review focuses on the emergence of porous crystalline framework catalysts for the thermocatalytic reduction of CO₂; future technologies for such transformations will be underpinned by the availability of low cost and sustainably sourced H₂.

Three principal conversion technologies exist for CO₂ reduction, namely photocatalysis, electrocatalysis and thermocatalysis, although variants such as photothermal,⁵⁶ electrothermal⁵⁷ and photoelectrocatalysis⁵⁸ have emerged. Each of these approaches has benefits and limitations, but they should not be viewed as competitors or alternatives since discoveries made in one often cross-fertilise advances in others. Thermocatalysis is suited for immediate adoption into existing CO₂ emission-intensive industries where H₂ is abundant and continuous product consumption guaranteed, as in petrochemical refining, and currently offers the highest reaction rates and product yields,⁵⁹ and hence suitability for scale-up. In contrast, electrocatalytic and photocatalytic CO₂ reduction are less energy intensive, typically operating at ambient temperature, and use water as a hydrogen source, but exhibit slower kinetics and usually produce a dilute aqueous/solvent product stream with high concomitant separation costs. Such technologies are therefore presently better suited to small scale, distributed CO₂ reduction, and require significant research and development to progress to demonstration plants wherein new challenges are expected. The levelized cost of organic products by electrocatalysis is currently high due to electrolyser capital costs, with cost per unit electrode area in turn dependent on their current density and faradaic efficiency.⁶⁰ Productivity is hampered by low gas solubility and poor diffusion (limited to a few tens of nm) through liquid electrolytes to the catalyst surface.⁶¹ Photocatalytic routes are currently hampered by inefficient light harvesting, charge separation and transfer to reactants, and slow multi-electron reduction kinetics resulting in low CO₂ conversion and poor quantum efficiencies. Photocatalytic^5,62–65 and electrocatalytic^66,67 CO₂ reduction over structured nanomaterials are the subject of comprehensive recent reviews, and hence this review focuses on the design, synthesis and application of solid catalysts derived from porous, crystalline frameworks for thermal CO₂ conversion.

Porous, crystalline frameworks are a broad class of crystalline materials which possess well-defined, ordered and two-dimensional (2D) or three-dimensional (3D) framework structures comprising specific secondary building units (SBUs) made of cations (or clusters) and organic anions such as metal organic frameworks (MOFs),⁶⁸ covalent organic frameworks (COFs)⁶⁹ and zeolitic imidazolate frameworks (ZIFs).⁷⁰ Their structures are usually visualised as nodes present at specific locations in 3D space, interconnected by linkers; a node is defined as a point in the framework where multiple linkers intersect. The diversity in composition, shape, size and nature of nodes, linkers and their connecting topology affords a plethora of possible crystalline frameworks. In 2017, the Cambridge Crystallographic Data Centre had catalogued 69 666 MOFs alone based on experimental crystal data.⁷¹

Research into crystalline frameworks has seen rapid growth, with numerous new applications of such materials as catalysts reported. Crystalline frameworks have received attention in photocatalysis^{56,66,72–78} electrocatalysis^66,76,79,80 and catalysis using thermal energy.^81,82 The rising interest in crystalline frameworks for thermocatalysis mirrors a deeper understanding of, and experimental improvements in, their stability at elevated temperatures and under reaction conditions. This is exemplified for MOFs, wherein the encapsulation of molecular catalysts and chemical fixation of CO₂ by liquid phase reactions has been reviewed,^82–87 along with some aspects of thermocatalytic CO₂ reduction.^81,88,89 However, the application of different crystalline framework classes, and their derivatives, as thermal catalysts for CO₂ reduction has not been the subject of a comprehensive and critical analysis to date.

Thermocatalytic CO₂ reduction is challenging in part due to the range of potential products, which may be thermodynamically favoured but with slow kinetics such as methanation (the Sabatier reaction) or methanol/dimethyl ether (DME) formation, or thermodynamically disfavoured as for partial reduction to CO. In general, a high CO₂ activation barrier necessitates high reaction temperatures, whereas H₂ activation requires high H₂ pressures; both decrease energy efficiency. There are also materials design challenges: CO₂ reduction requires molecular hydrogen to dissociate into H* adatoms proximate to an activated CO_x species, and hence the co-location of suitable active sites for each reactant. Porous, crystalline frameworks are an ideal platform to satisfy such spatial constraints due to their structural versatility. For example, metal nanoclusters encapsulated in a MOF can activate hydrogen, while linkers with appropriate functional groups can activate CO₂. Fig. 1a and b illustrate how small (8 × 14 atoms) capped copper clusters (hydrogen activation sites) could be incorporated in the tetrahedral voids of PCN-426(Ni) (PCN, Porous Coordination Network), proximate to the linkers. The distance between Cu atoms and linkers could be tuned by placing larger (6 × 62 atoms) Cu clusters into the octahedral voids of PCN-426(Ni), see Fig. 1c and d, enabling control over the separation between H₂ and CO₂ activation sites. Porous crystalline frameworks also permit careful design of the local environment surrounding catalytically active sites, acting as macroligands able to tune their steric and electronic properties, and hence reactivity.⁹⁰

Fig. 1 Two conceptual porous frameworks illustrating the encapsulation of different size (capped) copper nanoclusters within tetrahedral or octahedral cavities of PCN-426(Ni) with different Cu-linker separations. (a) Front view of 8 × 14 atom Cu cluster within tetrahedral voids of PCN-426(Ni). (b) Perspective view of structure in (a). (c) Front view of 6 × 62 atom Cu cluster within larger octahedral voids of PCN-426(Ni). (d) Perspective view of structure in (c). For illustrative purposes the capping agents are omitted. Orange: Cu; purple: Ni; yellow: O; green: C; white: H. These images were constructed using the open source molecular dynamics visualization software Ovito⁹⁹ and structural data for PCN-426(Ni)¹⁰⁰ and Cu nanoclusters;¹⁰¹ methodology is detailed in the ESI.†

Porous, crystalline frameworks are widely researched for selective gas sorption and molecular separations due to their high surface area, large pore volume, molecular sieving and tunable adsorbate–adsorbent interactions through surface functionalisation.⁹¹ Selective binding to framework metal atoms, differences in molecular packing efficiencies, framework flexibility and gate-opening mechanisms are effective strategies to facilitate molecular separation with enhanced selectivity.⁹² Many frameworks demonstrate a relatively high sorption capacity for CO₂ and H₂ arising from chemical interactions of these gases at nodes or linkers.^93,94 For example, Demessence et al. reported a post-synthetically modified, pendant alkylamine containing MOF, 1-DMF, with an isosteric heat of adsorption for CO₂ of 90 kJ mol⁻¹.⁹⁵ Lau et al. has shown that partial exchange of Zr ions with Ti ions in a UiO-66 framework (post-synthesis) doubles the CO₂ sorption capacity, mainly due to decrease in pore size and higher adsorption enthalpy in the resulting heterometallic Zr–Ti UiO-66 MOF.⁹⁶ Porous, crystalline frameworks offering enhanced H₂ and CO₂ sorption are highly desirable for thermocatalytic CO₂ reduction; as a linear molecule with a large ionisation potential and small electron affinity, CO₂ exhibits a high intrinsic activation barrier to reduction,⁹⁷ however frameworks possessing electron-donating ligands can activate CO₂ by charge transfer into its antibonding molecular orbitals.⁹³ Framework basicity is therefore regarded a prerequisite structural feature of CO₂ reduction catalysts. Using molecular simulations and comparison of a series of isoreticular MOFs (IRMOF1) differing in metal centre, organic linker, functional groups and framework topology, Babarao et al. have shown that CO₂ adsorption in porous, crystalline frameworks depends on their free volume and accessible surface area, and the presence of functional groups and type of linkers.⁹⁸ This review is organised as follows. Thermodynamic and process considerations of CO₂ reduction are first discussed in Section 2, followed by a concise discussion of the key reaction intermediates and governing mechanisms in Section 3. Strategies to leverage the design flexibility of catalysts derived from porous crystalline frameworks for CO₂ reduction are elaborated in Section 4. Critical consideration is given to the structural stability of crystalline frameworks under reaction conditions, often regarded as their greatest weakness in Section 5. Three broad approaches for the development of catalysts for thermal CO₂ reduction are identified: (i) nanoparticle encapsulation in porous crystalline frameworks; (ii) the use of crystalline frameworks as sacrificial precursors; and (iii) the integration of crystalline frameworks into composite catalysts. These approaches are elaborated in Sections 6, 7 and 8 respectively. A concise analysis of porous crystalline framework-derived catalyst deactivation is presented in Section 9, before conclusions and a discussion of future outlooks.

2.0 Thermodynamic & process considerations

While the thermodynamics of a reaction establishes limits on the reactant conversion and product selectivity, process parameters can be adjusted to kinetically control catalyst activity and selectivity towards desired products. In this section the thermodynamics of major CO₂ reduction reactions are analysed to identify ideal operating regimes with respect to process conditions. The choice of a specific porous, crystalline framework topology and composition for thermal catalysis should be guided by reaction parameters including temperature, pressure, space velocity, mechanical stress (on the catalyst), attrition resistance, and the physical/chemical properties of reactants, intermediates and products. After identifying a promising parameter set, a systematic search for a suitable crystalline framework is possible, based on its stability under the selected reaction environment. A reaction network for CO₂ reduction and associated reaction enthalpies is shown in Scheme 1, based on data for the reverse water gas shift (RWGS) reaction^102,103 and methanol synthesis^{5,6,104–108} wherein process conditions and catalyst performances have been compiled. RWGS being endothermic is favoured at high temperature, and hence CO₂ conversion increases with an increase in temperature (Fig. 2). The RWGS has been studied under various conditions and promising results reported for diverse systems. Lu et al. obtained 100% selectivity to CO and 40% CO₂ conversion over Ni/CeO₂ at 700 °C, 1 bar pressure and a H₂ : CO ratio of 1,¹⁰⁹ whereas Oshima et al. obtained the same conversion and selectivity at only 150 °C for a H₂ : CO ratio of 1 in the presence of an electric field.¹¹⁰ The former energy intensive conditions may render such a process uneconomical. Low conversion can be addressed by recycling the product stream (and unreacted CO₂) back into the reactor, and hence catalysts which are highly selective to CO even at modest conversions, but under mild process conditions, are preferable. The energy efficiency of processes integrated with RWGS can be improved in a hybrid mode by coupling electrochemical CO₂ reduction with downstream thermocatalytic process such as FT synthesis, although crystalline framework-derived catalysts are envisaged to offer promising low temperature RWGS activity due to the high dispersion of metal nanoparticles they facilitate. Selectivity control in RWGS is also attainable by varying the reactant space velocity, being favoured at high flow rates since CO formation involves fewer intermediates than for other products, with direct CO₂ dissociation, formate and carboxylate mechanisms proposed. However, a high space velocity can also induce rapid heat removal and hence lower CO₂ conversion, and such operation requires readily accessible active sites wherein rates are not constrained by reactant or product diffusion. Porous, crystalline frameworks usually satisfy these criteria due to their high surface area, pore volume and large pore apertures relative to the small molecules involved in RWGS (CO₂, CO and H₂O). Fixed bed, moving bed, fixed fluidised bed, fluidised bed and chemical looping reactors are all viable options for the RWGS reaction.

Scheme 1 CO₂ reduction reaction network and associated reaction enthalpies.

Fig. 2 Reverse water–gas shift reaction of CO₂ to CO as a function of H₂ : CO₂ ratio at different temperatures. Adapted from ref. 103 with permission from Elsevier, copyright 2017.

The enhanced dispersion and stabilisation of metal nanoparticles throughout framework-derived catalysts is also expected to promote activity for CO₂ methanation. Indeed, for MIL-101, incorporation of a 20 wt% loading of 2.5 nm Ni nanoparticles resulted in 100% CO₂ conversion with 100% selectivity to CH₄ at 300 °C, 10 bar and H₂ : CO ratio of 3.¹¹¹ Methanation is highly exothermic, and the thermodynamic conversion and selectivity both gradually decrease with increasing temperature, whereas higher pressures increase activity and selectivity (Fig. 3a and b). In exothermic reactions, heat management is critical for large-scale processes to limit catalyst deactivation by coking and/or sintering and to improve process economics through efficient use of waste heat.¹¹² A study on the impact of space velocity and cooling rates for CO₂ methanation using a transient mathematical model identified a negative impact of increasing reactant pressure due to faster catalyst coking induced by methane cracking.¹¹³ Methane yields improved with careful thermal management, with a complex, dynamic relationship observed between reactor performance, space velocity and heat removal rate. Heat exchanger type reactors are therefore best suited for large-scale methanation, although monolithic designs and certain scalable microreactors also enable efficient heat removal.¹¹⁴ At lab scale, fixed bed reactors have been favoured due to ease of operation. It is proposed that reactors and processing technologies developed for syngas (predominantly CO and H₂, with some CO₂) conversion could be adapted for CO₂ methanation due to the similar feed streams, with only minor engineering barriers anticipated.⁵⁹

Fig. 3 Thermodynamic equilibrium for CO₂ conversion to methane. (a) CO₂ conversion as a function of temperature and pressure. (b) CH₄ selectivity as a function of temperature and pressure. Adapted from ref. 108 with permission from Elsevier, copyright 2013.

Methanol formation from CO₂ is mildly exothermic reaction and favoured at low temperatures.¹⁰⁴ Thermodynamic conversion limits for H₂ : CO₂ ratios of 3 and 10 are shown in Fig. 4a and b, respectively (with corresponding methanol selectivities in Fig. 4c and d). For the former, at low (10 bar) pressure, CO₂ conversion decreases from 20% at 150 °C to only 14% at 200 °C, before slowly increasing slowly thereafter to a maximum of 30% at 350 °C. Under the same conditions, methanol selectivity decreases continuously from 95% at 150 °C to almost zero at 350 °C. High methanol yields therefore require catalysts able to activate CO₂ and H₂ <200 °C due to the sharp fall in thermodynamic selectivity to methanol at higher temperature. Above 250 °C there is minimal difference in CO₂ conversion (Fig. 4a) for 10 versus 30 bar, although significant improvement in methanol selectivity is obtained. However, relatively modest pressure increases do enhance CO₂ conversion at 150 °C, and methanol yield can therefore be significantly augmented by a moderate increase in pressure for catalysts with high activity at low temperature. There are very few reports of CO₂ reduction at such low temperatures. An increase in yield of CO or CH₄ is typically obtained when CO₂ reduction reaction takes place at temperatures higher than >250 °C.¹¹⁵ Since methanol synthesis from CO₂ is mildly exothermic, heat management is expected to play an important role (albeit less so than for methanation) in governing catalyst selectivity and stability. As for many continuous flow reactions, the space velocity should be tuned to maximise contact time with the catalyst while promoting desorption and removal of products from the reactor bed.¹⁰⁴

Fig. 4 Thermodynamic equilibrium curves for CO₂ hydrogenation to methanol. (a) CO₂ conversion and (b) methanol selectivity for a H₂ : CO₂ ratio of 3. (c) CO₂ conversion and (d) methanol selectivity for H₂ : CO₂ ratio of 10. Lines (i) to (vii) denote pressures of (i) 10, (ii) 30, (iii) 100, (iv) 200, (v) 300, (vi) 400 and (vii) 500 bar, respectively. Adapted from ref. 104 with permission from the American Chemical Society, copyright 2017.

Multi-tube fixed bed and fluidised bed reactors, microreactors and membrane reactors have all been utilised for methanol production from CO₂ at the lab scale.^116,117 Reactor and processing technology from syngas conversion to methanol can be adapted for CO₂ reduction. At the industrial level, three reactor types have been employed for methanol synthesis: (i) the tubular boiling water reactor by Lurgi and Haldor-Topsoe which uses boiling water to remove heat from catalyst loaded tubes to achieve an isothermal reaction regime and low catalyst sintering rates; (ii) the quench reactor developed by ICI Synetix (now Johnson Matthey) using adiabatic catalyst beds and a cold feed gas for quenching which offers significant economic advantage; and (iii) series adiabatic reactors developed by Haldor-Topsoe and Kellogg with low pressure drops and high methanol productivity.¹¹⁸ Copper on metal oxide supports such as ZnO and ZrO₂ are commonly employed for such engineering studies, however reaction parameter space should also be explored for new catalysts which exhibit high methanol selectivity and activity at low temperature, such as those derived from porous crystalline frameworks.

Recently, CO₂ conversions reaching 34% with high selectively (90%) to hydrocarbons (olefins and gasoline) were reported for Fe catalysts, termed CO₂-Fischer–Tropsch (CO₂-FT).³⁸ A tandem process for converting CO₂ to hydrocarbons is feasible since the reaction conditions are essentially identical to those for CO-FT, and some CO-FT catalysts also exhibit significant RWGS activity for in situ conversion of CO₂ to CO. Two challenges are identified for CO₂-FT: CO₂ can poison CO-FT catalysts; and water from the RWGS reaction can also deactivate CO-FT catalysts. Sustained production of hydrocarbons by CO₂-FT is therefore problematic in regard of catalyst design and reaction engineering. CO-FT has been investigated in fixed bed, fluidised bed and three phase slurry reactors. The latter have been the reactor of choice for CO-FT since they offer unique advantages in respect of efficient heat removal, enhanced heat and mass transfer between catalyst and reactants, and continuous operation.^119,120 For CO₂-FT, the removal of water and recycling of undesired products (CO, CH₄) back into the reactor must also be addressed. One solution is the use of independent modules, wherein a RWGS reactor feeds into a conventional CO-FT module, as this facilitates independent process optimisation for each reaction, which may prove intractable for a tandem process.⁵⁹

3.0 Mechanistic insights

Mechanistic studies provide atomic scale insight into reactions, enabling key interdependencies between active sites, intermediates and products to be identified, rarely possible from a macroscopic analysis of reaction data, and thereby guiding catalyst design. Mechanistic studies of electrocatalytic CO₂ reduction^121–123 have received more attention than those for photocatalytic^124,125 and thermocatalytic CO₂ reduction.^97,126,127 For thermocatalytic CO₂ reduction, four distinct mechanisms are reported. Formate (HCOO*) and RWGS pathways are the most commonly proposed (Fig. 5), however other pathways such as the hydrocarboxyl (trans-COOH*) and revised formate mechanisms have also been advanced and experimentally verified. The former two pathways are supported by the experimental observation of formate and carboxylate (COO*) surface intermediates,¹²⁸ although it is difficult to identify all reaction intermediates because of their transient lifetimes and limitations in the experimental techniques. The relative stabilisation of different reaction intermediates at catalyst surfaces dictates which reaction pathway, and hence product distribution, is favoured.

Fig. 5 Reaction pathways in thermocatalytic CO₂ reduction following formate, RWGS + CO and CO₂-FT routes from a common carboxylate intermediate. Adapted with permission from ref. 140 with permission from the American Chemical Society, copyright 2017.

3.1 Formate mechanism

The formate mechanism involves an initial hydrogen transfer to CO₂ to produce formate (HCOOH*) which can undergo sequential dissociation and reduction reactions to form dioxomethylene (H₂COO*), formaldehyde (H₂CO*), methoxy (CH₃O*) and methanol.¹²⁹ The detection of HCOO* species on Cu surfaces by XPS and TPD lends support to the formate mechanism.¹³⁰ Other studies report that low-coordinate ionic Cu species (such as Cu⁺) stabilise formate (HCOO*) and dioxomethylene (H₂COO*) intermediates promoting methanol formation,^131,132 and may constitute the active sites in Cu/ZnO catalyst. In the formate mechanism, CO produced by the RWGS reaction accumulates as a by-product of the reaction. Hydrogenation of CO generates an unstable formyl intermediate (HCO*) which tends to dissociate to CO and H. Methanol synthesis via the formate pathway is widely reported for Cu surfaces (Cu(100)¹³³), Cu clusters (Cu₂₉),¹³² Cu nanoparticles on metal oxides (Cu/ZnO,¹³² Cu/SiO₂,¹³⁴ Cu/ZrO₂¹³⁴), alloy surfaces (Zn deposited Cu(111),¹³⁵ Cu(100),¹³⁶ Cu(110)¹³⁶), metal carbides (Cu/β-Mo₂C)¹³⁷ and Pd based catalysts (Pd/β-Ga₂O₃,¹³⁸ Pd₄/In₂O₃¹³⁹). For the formate route, hydrogenation of the formate or dioxomethylene intermediate and C–O bond cleavage are proposed as the rate determining steps.

3.2 Revised formate mechanism

Cu/ZnO is used for the industrial conversion of syngas to methanol and considered the benchmark catalyst for CO₂ reduction to methanol. However, the origin of the high methanol selectivity for Cu/ZnO remains contentious. Recent studies over Cu(111) model catalysts suggest that HCOO* conversion to formic acid is more energetically favourable than to dioxomethylene,¹⁴¹ and a revised-formate (r-HCOO*) mechanism was consequently proposed wherein formic acid is the precursor to H₂CO* and eventually methanol.^97,141 Other studies support the revised-formate mechanism, and the high selectivity of Cu/ZnO and Cu/ZrO₂ catalysts to methanol can be rationalised based on the oxophilicity of metal oxides and their ability to stabilise reactive intermediates.^142–145 To date, the revised formate mechanism is proposed for Cu (ZnCu(211),¹⁴⁶ ZnO/Cu(111),¹⁴⁶ (Cu₄/Al₂O₃)¹⁴⁷) and Pd based systems ((Pd/CNT),¹⁴⁸ PdCu(111)¹⁴⁹).

3.3 RWGS + CO mechanism

The RWGS pathway involves CO₂ reduction to a carboxylate species (COOH*) which dissociates into OH and CO surface species.^140,150 Depending on the CO adsorption strength, it can either desorb or remain bound in an activated form (CO*). The catalyst and reaction conditions thereafter govern whether CO* is reduced to methane and/or methanol continuing on the RWGS pathway, or experiences C–O bond cleavage and subsequent reduction and C–C coupling following the CO₂-FT pathway to higher hydrocarbons. In the RWGS + CO reaction pathway, formate is considered a spectator species in methanol formation with COOH* favoured as the key reactive intermediate.¹⁵¹ Catalysts which strongly adsorb CO will facilitate methanol and hydrocarbon formation, whereas those which bind CO more weakly will favour the RWGS reaction and CO desorption. The low selectivity of pure Cu catalysts for CO₂ reduction to methanol is rationalised by the instability of formyl intermediate (HCO*) with respect to dissociation to CO.¹³² Dopants that stabilise formyl are therefore expected to increase selectivity to methanol, as predicted by density functional theory (DFT) calculations and Kinetic Monte Carlo (kMC) simulations for Ni/Cu(111) surface alloys.¹⁵² The RWGS + CO mechanism is proposed for Cu/CeO_x,¹⁵³ Cu/TiO₂,¹⁵⁴ Cu/ZrO₂¹⁵⁴ and Cu/ZnO/Al₂O₃¹⁵⁵ at very high pressures (360 bar). DFT and kMC simulations suggest a more significant role for the RWGS + CO mechanism over Pt, Rh, Ni, Pt doped Cu(111) surfaces.¹⁵²

3.4 The trans-COOH mechanism

Trace water, a by-product of the RWGS reaction, can influence activation barriers and the preferred reaction pathway,¹⁴³ switching on the trans-COOH* mechanism, wherein DFT calculations suggest that trans-COOH* formation, rather than formate, is rate-limiting over Cu(111).¹⁵⁶ Water lowers the energy barrier to trans-COOH* (and subsequent reactive intermediates) thereby improving selectivity to methanol. The trans-COOH mechanism proposes that hydrocarboxyl (COOH*) species are initially formed by reaction of adsorbed CO₂ with H atoms (provided by H₂O) to form COOH*. The COOH* is further hydrogenated to dihydroxycarbene (COHOH*), whose subsequent dissociation yields hydroxymethylidyne (COH*); hydrogenation of the latter then yields hydroxymethylene (HCOH) and hydroxymethyl (H₂COH) intermediates, and methanol.¹¹⁸ CO formation from CO₂ reduction may occur by either reduction of the COOH* intermediate and subsequent dissociation by the RWGS + CO pathway, or the direct dissociation of CO₂ to CO and O* by a redox mechanism.^143,157,158 The trans-COOH mechanism is suggested for metallic Cu,¹⁵⁶ Au/Cu–ZnO–Al₂O₃¹⁵⁹ and Ga₃Ni₅ surfaces.¹⁶⁰

4.0 Leveraging flexibility and versatility in the design of catalysts from crystalline frameworks

Complete and predictable control over physicochemical properties across length scales is the Holy Grail of materials science. While substantial progress has been made in fabricating nanoparticles,¹⁶¹ thin films¹⁶² and 3D solids,^163–165 generic multiscale materials design and synthesis remains beyond current modelling and fabrication capabilities. However, advances have been made in the development of single atom catalysts and nanoclusters stabilised on and within thin films,¹⁶⁶ metal oxides^167,168 and porous crystalline frameworks.^169,170 Reticular chemistry, defined as the linking of molecular building blocks into predetermined structures by strong bonds, has brought material science a step closer to the design of multiscale functional materials. Our contributions in this field span nanoporous films and 3D porous architectures for adsorption,^171,172 separation,^173–175 sensing,^176,177 reusable sensors,^178,179 antibacterial films^180–182 and catalysis.^{165,178,183–188}

Allendorf and Stavila draw an analogy between the hierarchical structure of porous crystalline frameworks composed of several sub-components, characterised by different length scales and energetics, and interprotein assemblies (Fig. 6).¹⁸⁹ Indeed such frameworks can be visualised and understood at several hierarchical levels, ranging from metal ion–ligand interactions, to host–guest interactions and interpenetration. In porous crystalline frameworks, nodes, typically comprising multivalent metal ions or metal clusters, are connected by organic ligands (linkers) to form 3D networks. The treatment of such frameworks as structurally ordered solids composed of secondary buildings units (SBU), i.e. nodes and linkers, aids their visualisation and the exploration of novel node and linker combinations, facilitating the design of new materials using a mix-and-match approach. The structural diversity achievable by this approach is evidenced by the vast number of such frameworks known, and accelerating discovery of new structures. Such diversity is enabled by composition independent, and charge density governed, interactions of nodes with linkers, allowing the exchange of nodes and linkers with similar charge densities.¹⁹⁰ As previously noted, these porous crystalline frameworks provide high surface areas and pore volumes, selective gas adsorption, molecular confinement and excellent dispersion of active sites.^191,192 Considering CO₂ reduction, catalysts derived from porous crystalline frameworks should possess three key features: (i) high CO₂ and H₂ sorption capacities; (ii) a high density of active sites for efficient CO₂ and H₂ activation ideally co-located to promote transport of reactive intermediates; and (iii) a low affinity for the desired product (whether CO, CH₃OH, CH₄ and/or H₂O). The rational design and synthesis of functional crystalline frameworks has been extensively reviewed.^86,190–198 Their physicochemical properties can be modulated to a large extent through judicious selection of appropriate linkers, nodes and resulting cavity volumes.

Fig. 6 Comparison of the structural hierarchy of porous crystalline frameworks and interprotein assemblies. Reproduced from ref. 189 with permission from The Royal Society of Chemistry, copyright 2015.

Active sites can be incorporated in the form of coordinatively unsaturated metal sites, or the stabilisation of single atoms within the crystalline framework. Activities of individual catalytic sites in multifunctional catalysts derived from frameworks can also be modulated to tune overall activity and product selectivity. Finally, the shape, size and porosity of crystalline materials can be engineered to enhance molecular transport (overcoming potential molecular diffusion limitations^199,200) and tune the hydrophobicity of active sites (to displace reaction products). These aspects of catalyst design are summarised below.

4.1 Linker and node modification in crystalline frameworks

Linkers in porous crystalline frameworks can be modified during synthesis or post-synthetically to incorporate metal or organic functions. Examples of metallated linkers (linkers containing metal moieties) used in framework synthesis^201,202 and post-synthetic strategies²⁰³ are shown in Fig. 7a and b respectively. Linkers can also be modified to incorporate new organic functions²⁰⁴ (Fig. 7c).

Fig. 7 Strategies to modify linkers in porous crystalline frameworks. (a) Use of metallated moieties such as a metalloporphyrin tetracarboxylic ligand. (b) Post-synthetic metalation of linkers incorporated into the framework. (c) Post-synthetic functionalisation of linkers with organic moieties. (a) Reproduced from ref. 202 with permission from The Royal Society of Chemistry, copyright 2016; (b) reproduced from ref. 203 with permission from The Royal Society of Chemistry, copyright 2015; (c) adapted from ref. 203 and 204 with permission from The Royal Society of Chemistry, copyright 2015.

Covalent Organic Frameworks (COFs) are porous, organic two-dimensional or 3D solids composed of light elements (H, B, C, N, O) linked by strong covalent bonds. COFs have recently been synthesised with single active sites using porphyrin ligands containing Fe,^205,206 Cu,^207,208 Ni,²⁰⁷ Co^201,202 and Mn^206,209 centres. The high thermal stability of COFs, and their functionalisation by metallic moieties, has made them attractive potential catalysts. Metalloporphyrins can introduce Lewis acidity into frameworks through modulating the electron density.¹⁹⁰ Phthalocyanines,²⁰⁵ bipyridine,²¹⁰ phenylpyridine,²¹¹ N-heterocyclic carbine gold(I)²¹² and salen complexes^213,214 are also reported as molecular linkers with which to incorporate metals into crystalline frameworks for catalytic applications. Nodes in porous crystalline frameworks can also be modified by either bottom-up approaches (using metal ions or clusters) or post-synthetic cation substitution or the attachment of single sites^190,215,216 as shown in Fig. 8a, b, c and d, respectively.

Fig. 8 Strategies for node modification in crystalline frameworks. (a) Use of metal ions or clusters as nodes assembled using post-synthetic modification. (b) Bottom-up synthesis by cation substitution. (c) Post-synthetic attachment of single metal sites to nodes. (d) Node transformation in PCN-700(Zr) involving conversion of Zr₆ clusters to Zr₆Ni₄ clusters by solution phase grafting referred to as solvent-assisted incorporation of metals (SIM). (a and b) Adapted from ref. 190 with permission from Wiley-VCH, copyright 2018; (c) adapted from ref. 215 with permission from the American Chemical Society, copyright 2013. (d) Adapted from ref. 216 with permission from Wiley-VCH, copyright 2015.

4.2 Cavity modification in crystalline frameworks

The chemical composition of the inner surface of framework cavities (pores) and cavity dimensions directly impact molecular adsorption and hence catalysis. The shape, size and chemical characteristics of cavities in crystalline frameworks can be varied by the modification of linkers or nodes, encapsulation of nanoclusters or novel method such as pore space partitioning. Pore space partitioning uses the shape of linkers to partition a large cavity into several smaller ones. Grafting of an appropriate molecular function to a partitioning linker enables repatterning of the parent framework (Fig. 9a and b). The pore space partitioning approach has been used to reduce the cavity of FJU-88 (12.0 Å × 9.4 Å) through introducing a triangular ligand, resulting in a new MOF FJU-90 with a cavity size of 5.4 Å × 5.1 Å offering improved separation of acetylene from CO₂.²¹⁷ Symmetry and size-matching pore partitioning agents such as 2,4,6-tri(4-pyridinyl)-1,3,5-triazine or a metal-complex cluster have also been introduced to the hexagonal channels of MIL-88/MOF-235 resulting in dramatically enhanced ammonia sorption.²¹⁸ The size, shape and chemical characteristics of nanoporous cavities wherein catalysis predominantly occurs, can hence be readily tuned in crystalline frameworks.

Fig. 9 Pore space partitioning as a strategy for cavity modification. (a) Conversion of FJU-88 to FJU-90, seen along the crystallographic c-axis. (b) Polyhedral representation of FJU-88 conversion to FJU-90. Adapted with permission from ref. 217 with permission from the American Chemical Society, copyright 2019.

4.3 Single atom/site catalysts

Isolated metal sites at the corners, edges and facets of catalytic nanoparticles play an important role in governing their activity and selectivity. For example, a direct correlation between product selectivity in CO₂ reduction and active site density is reported for Rh nanoparticles supported on TiO₂.²¹⁹ The densities of isolated (Rh_iso) and total (Rh_NP) surface sites were estimated from diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) of adsorbed CO. Turnover frequencies (TOFs) for the RWGS reaction (producing CO from CO₂) strongly correlated with Rh_iso sites, whereas those for CH₄ strongly correlated with Rh_NP sites. Reaction-induced restructuring altered the relative distribution of these sites resulting in changing product selectivity with time on stream.²¹⁹ Engineering strategies such as varying the size (and size distribution) and morphology of nanoparticles/crystals have also been used to optimise the density and uniformity of isolated active sites. Homogeneous catalysis has drawn inspiration from biocatalysis to design molecular catalysts with single active sites; although this approach ensures extreme uniformity in catalysts it cannot circumvent shortcomings associated with product separation and continuous operation. The design of heterogeneous catalysts from porous crystalline framework containing single active sites (which may be individual atoms or ensembles thereof) has therefore been enthusiastically pursued.^220,221 Single sites are defined as isolated, well-defined, spatially separated and, ideally, structurally identical species,²²⁰ and when present in heterogeneous catalysts promote shape selectivity and specific interactions.²²⁰ Single active sites can be incorporated into crystalline frameworks using coordinatively unsaturated nodes²²² or metalation of linkers (as described in Section 4.1). Single atom catalysts (SACs) are a sub-class of single site catalysts in which individual site-isolated (usually metal) atoms are dispersed on a support.^167,168,223 Single site (molecular) catalysts can be distinguished from SACs based on their coordination, since single sites have more coordination bonds with functional groups of the framework whereas single atoms are largely uncoordinated and dangling from a few anchoring bonds. SACs exhibit unique catalytic properties because of their low coordination, finding applications in hydrogenation, oxidation and organic synthesis. Crystalline frameworks have been used to create SACs by anchoring metal atoms at defect¹⁶⁹ and coordinatively unsaturated sites,¹⁷⁰ or pyrolysis of metal-functionalised framework to yield heteroatom-functionalised carbons with metal single atoms embedded.²²⁴

4.4 Diffusion limitations

Molecular diffusion of reactants and products is especially important for nanoporous framework-derived catalysts wherein mass-transport to/from active sites can be rate-limiting.⁸¹ For CO₂ reduction, special attention must be given to the relative diffusion rates of H₂ and CO₂ reactants in the nanopores of crystalline framework derived catalysts. These rates may differ due to differences in molecular size and/or solubility in liquid-filled pores, resulting in exposure of active sites to a different local H₂ : CO₂ molar ratio than in the bulk feed stream, impacting on reduction activity and selectivity. Experimental studies on CO versus H₂ are well documented for the Fischer–Tropsch reaction. Post et al. identified that the limited mobility of reactant molecules, especially H₂, in liquid-filled pores was a major limitation for synthesis gas conversion to fuels via the FT process over iron and cobalt catalysts in a fixed bed reactor.²²⁵ Catalyst activity may therefore be better optimised by considering the likely in-pore reactant ratios, rather than the bulk ratios typically considered for process control. Vervloet et al. attempted to identify favourable process conditions with respect to intra-particle diffusion limitations to maximise C₅₊ productivity for cobalt FT catalysts.²²⁶ They observed the H₂ : CO ratio increased towards the centre of catalyst particles, hindering chain growth probability due to intrinsically imbalanced H₂ and CO diffusivities and consumption. Interestingly, C₅₊ space time yield could be increased 3–10 fold by lowering the bulk H₂ : CO ratio from 2 to 1 and increasing the reaction temperature by 30 K.²²⁶ Differential diffusion is expected to be even more significant for CO₂ reduction due to its greater molecular size/mass relative to H₂. Yao et al. investigated H₂/CO/CO₂ gas mixtures for the FT reaction over cobalt and iron catalysts and observed that CO₂-rich reactant mixtures produced paraffin-rich products, whereas CO-rich reactant mixtures produced olefin-rich products.²²⁷

There is only limited literature on diffusion limitations for crystalline framework derived catalysts. Wang et al. compared the performance of Pt@ZIF-8 catalysts with different crystallite sizes for cylcooctene and hexene hydrogenation, observing that the product distribution depended on both pore and crystal sizes (Fig. 10).²²⁸ Larger cyclooctene molecules could not access framework pores due to their small openings, and hence hydrogenation was suppressed in all cases. In contrast, the n-hexane yield was inversely proportional to crystallite size, with the smallest (40 nm) particles offering the least diffusional resistance.

Fig. 10 (a) Illustration of differences in path travelled by reactants in crystals of different sizes (b) Differences in product distribution obtained for cyclooctene and hexene hydrogenation with Pt@ZIF-8 catalysts of different particle size. Reproduced from ref. 228 with permission from Springer Nature, copyright 2017.

Compared to their porous inorganic analogues (e.g. mesoporous silicas and carbons), porous crystalline framework catalysts are expected to exhibit better mass transport properties arising from their:

(1) Higher surface areas and larger pore volumes. Crystalline frameworks can also be engineered to relieve diffusion constraints by introducing pore hierarchy (hollow crystals, three-dimensional ordered microcrystal (3DOM) and/or mesopore incorporation) and pore aperture engineering.^229,230

(2) Availability of ordered sub-nanometer pore channels, which can facilitate molecular sieving to optimise the local concentration of reactant molecules at active sites. Framework flexibility and gate-opening mechanisms in crystalline frameworks as well as strategies like pore space partitioning also have potential for contribution to optimisation of the local environment.

(3) Tunable chemical functionality, which enables adsorbate–adsorbent interactions to be carefully optimised through pore surface, linker and node functionalisation and modification to retard or accelerate in-pore molecular transport.

4.5 Multifunctional catalysts based on crystalline frameworks

The interplay between different components of a multifunctional catalyst can also influence catalytic reactivity. For example, synergy between acid and metal functions is critical to optimise isomer production in bifunctional n-hexane hydroisomerisation²³¹ and anisole hydrodeoxygenation.²³² Similarly, the activity and density of hydrogen activation sites must be balanced against those of CO₂ activation sites, with these being co-located to avoid diffusional barriers, for an efficient CO₂ reduction catalyst. For example, high methanol selectivity (80%) and yields (297 mg g_cat⁻¹ h⁻¹) were obtained over a Cu–ZnO–ZrO₂ catalyst at 220 °C, 30 bar pressure and H₂:CO₂ 3.1.²³³ ZnO–ZrO₂ interfaces were identified as the active sites for CO₂ adsorption and activation, whereas metallic Cu species were identified to be the active sites for hydrogen activation and dissociation. The density, activity and spatial proximity of active sites in the catalyst was controlled by systematic changes in catalyst composition and porosity (using colloidal crystal templating), which led to synergetic multifunctional catalysis.²³³ The design flexibility of porous crystalline framework presents new opportunities to regulate interdependent interactions between different active sites. For example, synergetic multifunctional catalysis between Cu, ZnO and ZrO₂ components could be leveraged by dispersing ultrasmall Cu/ZnO_x nanoparticles in UiO-66-bpy MOF, to obtain complete methanol selectivity and stable catalytic activity up to 100 hours, albeit at lower yields (136 mg g_cat⁻¹ h⁻¹) to the above example, at 250 °C, 40 bar pressure and H₂:CO₂ of 3.²³⁴ High selectivity and catalyst stability are more desirable, since low CO₂ conversion can be alleviated using product recycle.

4.6 Product release from catalysts based on crystalline frameworks

The release of products from CO₂ reduction can be aided by the hydrophobisation of catalysts, or their construction as membranes. Methanol and water are major products of CO₂ reduction (irrespective of the pathway followed), and their displacement from crystalline framework-derived catalysts can be assisted by the introduction of hydrophobic moieties adjacent to active sites; this can enhance catalyst stability and activity. Aguado et al. investigated a thin shell of Substituted Immidazolate Material (SIM) on alumina beads as a catalyst for Knoevenagel condensation, and measured intrinsic reaction rates free from external and internal diffusional limitations.²³⁵ A direct correlation was observed between the contact angle of water measured for a thin film of SIM on a flat surface and corresponding catalytic activity for Knoevenagel condensation of benzaldehyde and malononitrile wherein a seven-fold rate enhancement was reported. Recently, Li et al. demonstrated the use of a Na⁺-gated water conducting membrane, composed of a defect free film of zeolite NaA, grown on ceramic hollow-fibers, to efficiently separate water from the gaseous reactants and products of CO₂ reduction; H₂O:CO₂, H₂O:H₂, H₂O:CO and H₂O:MeOH selectivities of 551, 190, 170 and 80 were respectively observed.²³⁶ CO₂ reduction was subsequently investigated for a Cu/ZnO/Al₂O₃ catalyst loaded on the outer membrane surface, with the water permeate collected from the membrane core. Catalytic CO₂ conversion and methanol yields were >2.6 times and >2.8 times higher respectively with than without the membrane at comparable selectivity (95% methanol in liquid products).²³⁶ Efficient molecular sieving between water and gas molecules >200 °C and >20 bar arose from the absence of defects and gating effect of Na⁺ located in 8-oxygen ring apertures of zeolite NaA, which regulated their effect size (Fig. 11). Such results for a commercial catalyst are unprecedented and highlight the importance of water removal for promoting CO₂ reduction.

Fig. 11 Application of a water-conducting membrane in CO₂ reduction. (a) Schematic of water-conducting membrane dehydration reactor for high purity methanol production. (b) Catalyst performance for 100 h on-stream at 220 °C and 35 bar, H₂ : CO₂ ratio of 3 and GHSV of 4200 ml g_cat⁻¹ h⁻¹ for a membrane length of 45 mm. (c) Different membrane preparation routes: (i) 50 to 200 nm seeds of NaA are attached to ceramic hollow-fibers by reaction with surface hydroxyls and annealing treatment to form a defect free membrane. Omitting the attachment step results in a defective membrane coating and poor selectivity. Decreasing the density of NaA seeds by a factor of 2 or 10, or using 300–400 nm and 400–700 nm NaA seeds also leads induces defects; (ii) a defect-free membrane effectively sieves water from gas molecules; (iii) defective membranes do not selectively transport water. Reproduced from ref. 236 with permission from Science, copyright 2020.

Incorporation of hydrophobic functions is also reported to stabilise crystalline frameworks under hydrous conditions during CO₂ reduction. Omary et al. developed a superhydrophobic fluorine-based MOF which exhibited excellent stability in water.²³⁷ Methyl groups in ZIF-68, ZIF-69 and ZIF-70 also exhibit good stability in boiling water, methanol, benzene and even 8 M NaOH for up to 7 days, due to suppressed hydrolysis of [ZnN₄] units associated with hydrophobic framework pores.^238,239 Pyrazolate-based Ni₈L₆ MOF,²⁴⁰ BUT-12 and BUT-13²⁴¹ likewise demonstrate good stability against water, acids and bases due to their hydrophobicity. Hydrophobisation can also be achieved by post-synthetic modification of crystalline frameworks using covalent,²⁴² dative²¹⁰ and click chemistry²⁴³ approaches.

5.0 Crystalline framework stability as a critical design parameter

The historical poor thermochemical stability of crystalline frameworks has previously limited their use in secondary processing and practical applications at industrial scale. For example, MOF-5 showed promise for several applications^244,245 but could not be practically implemented due to gradual decomposition by moisture in air.²⁴⁶ However, due to the advent of new synthetic and modification strategies, and progress in theoretical understanding of bonding in crystalline frameworks,²⁴⁷ such limitations have been overcome. Consequently, crystalline frameworks now find application in areas such as the detection and removal of antibiotics (Nitrofurazone and Nitrofurantoin) and organic explosives (2,4,6-trinitrophenol and 4-nitrophenol) in water.²⁴¹ The possibility of hydrothermal synthesis, in addition to solvothermal synthesis, has also improved the green credentials for crystalline framework production,^197,248 and their resulting industrial scale-up.²⁴⁹ The stability of crystalline frameworks has been extensively reviewed,^238,250,251 hence we focus here only on stability issues pertinent to thermocatalytic CO₂ reduction. Current understanding suggests that the relatively labile nature of coordination bonds that underpin most crystalline framework is responsible for their poor thermochemical stability.²⁵⁰ In designing a stable crystalline framework, the first factor for consideration is the node–linker bond strength, which is governed by charge density matching of linkers and metal centres (nodes). Using Pearson's hard/soft, acid/base (HSAB) principle, Zhou et al. identified two types of stable crystalline frameworks: those based on hard carboxylate linkers coordinated to hard, high valent metal ions; and those based on soft azolate linkers coordinated to soft, low valent metal ions.²³⁸ MIL-53,²⁵² MIL-100,²⁵³ MIL-101²⁵⁴ and Zr⁴⁺ based MOFs²⁵⁵ belong to the former hard base–hard acid type and are reportedly stable in water and even acidic conditions. However, crystalline frameworks of this type readily decomposed in basic media, as observed for PCN-222²⁵⁶ and MOF-545²⁵⁷ comprising Zr⁴⁺ and carboxylate linkers. Their poor stability in basic media was attributed to the competition of hard hydroxide anions with hard carboxylate linkers for coordination to metal cations. Carboxylate linkers have a relatively low affinity for protons due to their low pK_a values and form strong coordination bonds with metal ions, leading to a good stability in acidic media.²³⁸ However, hydroxide anions have a stronger affinity for metal ions than carboxylate linkers, resulting in ligand exchange and framework degradation. Zeolite Immidazolate Frameworks (ZIF),²³⁹ PCN-601,²⁵⁸ triazole⁹⁵ and pyrazolate²⁵⁹ based MOFs which belong to the latter soft acid–soft base group exhibit excellent stability in water and basic media, but readily decompose in acidic media. The high pK_a values of azolate linkers make them particularly reactive with protons leading to degradation of metal–linker bonds. In addition to protons and hydroxide ions, other hard anions such as F⁻, CO₃²⁻ and PO₄³⁻ can also degrade crystalline frameworks,²³⁸ and their presence can be mitigated by synthesising frameworks based on soft metal ions and soft linkers. For example, PCN-602, comprising soft metal ions and linkers, exhibits excellent stability in 1 M KF, K₃PO₄ and Na₂CO₃ whereas PCN-224 (assembled from Zr₆(OH)₈ nodes and tetrakis(4-carboxyphenyl)porphyrin linkers) completely degraded under such conditions.²⁶⁰ Pearson's HSAB principle provides a useful rule of thumb for predicting the stability of crystalline frameworks.

Structural connectivity and rigidity also influence the stability of porous crystalline frameworks. High connectivity through multiple bonds can stabilise frameworks even if some bonds are weakened or degraded. Linker rigidity imparts stability to UiO MOFs (UiO-66, UiO-67 and UiO-68) constructed from a homologous series of linkers.²⁵⁷ Greater flexibility of the longer linkers in UiO-68 and UiO-67 reduced their stability compared to UiO-66.

For CO₂ reduction to methanol, crystalline frameworks should be stable at temperatures up to 250 °C and in the presence of water vapour. Although crystalline frameworks are stable to water under ambient conditions, their stability under high partial pressures of water and methanol at reaction temperature has not been investigated. Heterogeneous catalysts containing metal functions typically require activation by reductive pretreatments at temperatures of 300–400 °C, and hence catalysts derived from crystalline frameworks should ideally by stable up to 400 °C under a reducing atmosphere. This is important for RWGS catalysts wherein frameworks must be stable towards both water vapour and CO (itself a reductant). Catalysts which favour formic^261,262 or acetic acid²⁶³ formation are most likely to be stable when based on crystalline frameworks constructed from hard acids/hard bases.

6.0 Nanoparticle encapsulation in crystalline frameworks

Nanoparticle encapsulation in porous crystalline frameworks is an effective strategy for preparing multifunctional catalysts.^264–266 Crystalline frameworks enable the encapsulation of extremely small nanoparticles which are stabilised by interactions with functional groups in linkers. Encapsulated nanoparticles may have higher catalytic activities than their naked counterparts, despite their confinement in framework cavities and coordination to linkers,²⁶⁷ attributed to strong metal support interactions and/or metal–ligand charge transfer. The surrounding framework can also induce shape and size selective catalysis for confined nanoparticles, wherein the geometric and electronic environment of reactants/products within cavities can be tuned by modifying the nodes and linkers,²⁶⁸ and where entropic considerations may become rate-controlling.²⁶⁹ Encapsulation of nanoparticles within crystalline frameworks also facilitates cooperative catalysis, wherein distinct active sites may participate in simultaneous (often simply termed bifunctional or one-pot) or sequential (termed tandem or cascade) catalytic reactions.²⁷⁰ This section describes the application of nanoparticles encapsulated within crystalline frameworks for CO₂ reduction on the basis of the major product.

6.1 Methanol selective CO₂ reduction catalysts

An early use of crystalline frameworks as catalysts can be traced back to 2008 when Müller et al. investigated Cu/ZnO supported on MOF-5 for CO₂ reduction.²⁷¹ Volatile precursors [CpCuL] (L = PMe₃, CN^tBu) and ZnEt₂ were incorporated in the MOF matrix by solventless, gas phase adsorption and decomposed by hydrogenolysis or photoassisted thermolysis (Cu) and annealing (ZnO) to obtain Cu/CuO/MOF-5, ZnO/MOF-5 and Cu/ZnO/MOF-5 catalysts. The Cu/ZnO/MOF-5 (1.4% Cu, 40.1% ZnO) catalyst achieved a methanol yield 60% that of the industrial catalyst (60% Cu/30% ZnO) and productivity of 212 μmol g_cat⁻¹ h⁻¹ despite the far lower Cu loading. This excellent performance was attributed to a high dispersion of Cu nanoparticles on the MOF support, and Cu/ZnO interfacial interaction. However, the catalyst degraded after few hours on-stream due to MOF-5 deactivation by water vapour. In 2016, Rungtaweevoranit et al. achieved an eight-fold enhancement in methanol yield (at 100% selectivity), compared to a commercial Cu/ZnO/Al₂O₃ catalyst, using 1% Cu/UiO-66 which comprised 18 nm copper nanoparticles individually encapsulated within UiO-66 crystallites (note these particles are larger than the framework cavities).²⁶⁷ Polyvinylpyrrolidone (PVP) protected copper nanoparticles were added to MOF precursors during the catalyst synthesis. Of the controls examined (Cu/MCF-26, Cu/ZrO₂, Cu/Al₂O₃, Cu/MIL-101, Cu/ZIF-8, and Cu/ZnO/Al₂O₃), methanol formation was only observed for Cu/ZrO₂, Cu/ZnO/Al₂O₃ and Cu/UiO-66, highlighting the importance of Cu–Zn and Cu–Zr interactions in methanol formation. The activity of encapsulated Cu nanoparticles was twice that of Cu nanoparticles impregnated post-synthesis of the MOF, and exhibited superior stability, highlighting the benefits of encapsulation. Cu/UiO-66 catalyst outperformed Cu/ZnO/Al₂O₃ at all temperatures, and was completely selective to CH₃OH whereas the commercial catalyst formed CO >200 °C. XPS evidenced partial reduction of Zr for the impregnated Cu/UiO-66 catalyst, likened to the strong metal support interaction (SMSI) observed for metal nanoparticles on reducible oxides.²⁷² An et al. also investigated CuO/ZnO nanoparticles on a UiO-bpy MOF, reporting 100% methanol selectivity and a space time yield of 2.59 g_MeOH kg_Cu⁻¹ h⁻¹ for 100 hours at 250 °C and 40 bar.²³⁴ CuCl₂ and ZnEt₂ were used to introduce Cu and Zn (Fig. 12) to the MOF. Cu binding to bpy ligands was verified by UV-VIS-NIR and the observation of a metal–ligand charge-transfer (MLCT) band at 470–530 nm. This MCLT band disappeared after catalyst reduction at 250 °C and 40 bar H₂, coincident with a copper reduction peak (Cu²⁺–Cu⁰) in temperature programmed reduction (TPR) studies, confirming a strong interaction between Cu nanoparticles and the MOF framework. XPS evidenced Cu⁰, Zn²⁺, Zn⁰, Zr⁴⁺ and Zr³⁺ species in the as-prepared catalyst. A related Cu/UiO-bpy catalyst was more active (6%) than CuZn/UiO-bpy catalyst (3%) but showed worse methanol selectivity (52%). Increasing the Zn : Cu ratio from 0.26 to 0.91 increased the methanol space time yield from 1.05 to 2.61 g_MeOH kg_Cu⁻¹ h⁻¹, emphasising the importance of Cu–Zn interactions in methanol formation as reported for conventional bimetallic catalysts.^5,104,105 The bpy ligands and Zr oxy nodes were also important for good catalyst activity, since a poorer performance was obtained for a related Zr-based MOFs with a different linker (UiO-67) and an Al-based MOF-253. Mechanistically, it was proposed that H₂ was first activated on Cu sites and subsequently spills over to O-deficient ZnO_x and Zr sites at which CO₂ is activated (to form carbonate and bicarbonate) and then reduced to methanol (Fig. 13). Kobayashi et al. compared the performance of Cu/γ-Al₂O₃ (Al³⁺), Cu/ZIF-8 (Zn²⁺), Cu/MIL-100 (Cr³⁺), Cu/UiO-66 (Zr⁴⁺), Cu/UiO-66-COOH (Zr⁴⁺), Cu/UiO-66-NH₂ (Zr⁴⁺) and Cu/Hf-UiO-66 (Hf⁴⁺) for CO₂ reduction at 220 °C, 2 atm and a H₂ : CO₂ ratio of 5 (Fig. 14a). They also proposed an interfacial charge-transfer between Cu atoms and the UiO-66 framework, correlating Zr/Hf binding energy shifts and methanol productivity (Fig. 14b).²⁷³ Methanol formation over Cu/UiO-66 was 70 times higher than for Cu/γ-Al₂O₃, with a further three-fold enhancement observed for hafnium analogues (Cu/Hf-UiO-66). Amine functionalisation of Cu/UiO-66 (Zr⁴⁺) led to marginal decrease in catalytic performance, however functionalisation with electron-withdrawing carboxylic acid groups increased methanol productivity three-fold, further evidencing that charge-transfer from Cu nanoparticles to the MOF framework was responsible for promoting CO₂ reduction to methanol versus Cu/γ-Al₂O₃.

Fig. 12 Synthesis of a CuZn@UiO-bpy catalyst for selective CO₂ reduction to methanol. Adapted from ref. 234 with permission from the American Chemical Society, copyright 2017.

Fig. 13 Cooperative catalysis using crystalline frameworks. (a) Methanol space-time yield versus reaction time. (b) Methanol selectivity versus time on-stream (c) Proposed mechanism for CO₂ hydrogenation at Cu/Zn/Zr interfacial sites. Adapted from ref. 234 with permission from the American Chemical Society, copyright 2017.

Fig. 14 (a) Methanol productivity for crystalline framework encapsulated Cu nanoparticles. (b) Correlation between methanol productivity and binding energy shift of support metal component following Cu encapsulation. Adapted from ref. 273 with permission from The Royal Society of Chemistry, copyright 2019.

The preceding examples highlight the importance of active sites arising from electronic interactions between metallic Cu and ZnO or ZrO_x nodes in crystalline frameworks (UiO-66 and UiO-66-bpy). Although a variety of other crystalline frameworks have been investigated for methanol formation, these have proved unsuccessful. Note that such electronic perturbation of Cu appears a consequence of framework confinement rather than direct bond formation, and that electron-donating and withdrawing framework linkers offer fine control over the electronic state of Cu nanoparticles. This contrasts with conventional supported catalysts which only offer limited control over the local geometric and electronic environment around metal active sites. Although reports of methanol selective catalysts remain limited, spatial confinement and localised electronic perturbation of metal clusters/nanoparticles to modulate their environment and reactivity is an emerging paradigm. A performance comparison of porous crystalline framework derived methanol, ethanol and formic acid selective catalysts prepared by various routes appears in Table S2 (ESI†).

6.2 Methane selective CO₂ reduction catalysts

A range of metals have been studied for the Sabatier reaction, including Pt, Pd, Rh, Ru, Fe, Co and Cu, however Ni nanoparticles supported on inorganic metal oxides such as Al₂O₃, TiO₂, SiO₂, ZrO₂, CeO₂, MgO and CaO have emerged as the most promising methanation catalysts²⁷⁴ due to their relatively high activity and Earth-abundance/low cost. Ni catalysts for CO₂ methanation based on encapsulated nanoparticles within crystalline frameworks (MIL-101) have been developed which achieve complete conversion and 100% methane selectivity for up to 200 hours.¹¹¹ Nanoparticle size and reducibility were identified as critical parameters that control activity and selectivity, being a function of the catalyst preparation and nature of the encapsulating framework (and resulting Ni interactions).

Zhen et al. investigated Ni-MOF-5 for the Sabatier reaction at 1 bar and a H₂ : CO₂ ratio of 4.²⁷⁵ CO₂ conversion increased with Ni loading, reaching a plateau ∼10 wt%, and was completely selective to methane at all conversions. A stable conversion of 75% was obtained at 320 °C for 80 hours, after which CO production was also observed. The high activity of this catalyst was ascribed to small Ni particles (9 nm) and a corresponding high metal dispersion (42%) and the facile reduction of Ni at 295 °C observed by TPR; a 10 wt% Ni/SiO₂ catalyst with larger particles (42 nm) showed little reduction at this temperature. Readily reducible Ni species, associated with a SMSI, are likely active species for CO₂ methanation. CO₂ reduction was postulated to occur by initial reduction to CO and subsequent reduction to methane on Ni(111) facets. Ni/UiO-66 was also explored for this reaction.²⁷⁶ Ultra-dispersed Ni nanoparticles (2 nm) throughout the high surface area UiO-66 framework, and associated SMSI, increased the TOF to 155 h⁻¹ at 320 °C and 68% conversion compared to 94 h⁻¹ for Ni-MOF-5 at the same temperature and 75% conversion; both 100% selective to methane. Ni/UiO-66 was stable for 100 hours on-stream. However, the best results for CO₂ methanation were obtained using 20 wt% Ni/MIL-101 prepared by double solvent impregnation.¹¹¹ At 10 bar and a H₂ : CO₂ ratio of 4, this delivered complete CO₂ conversion and 100% methane selectivity at 300 °C, with a TOF of 188 h⁻¹ stable for up to 200 hours. The high CH₄ productivity was attributed to a high Ni dispersion (2.5 nm particles) and concomitant strong SMSI, which lowered the Ni(O_x) reduction temperature to only 252 °C, and a high proportion of Ni(111) facets from the double solvent synthesis. DFT calculations suggest that the activation barrier for CO₂ dissociation over Ni(111) facets of 10 kcal mol⁻¹ is half that over Ni(200) facets. Mihet et al. compared the performance of MIL-101 and UiO-66 encapsulated Ni nanoparticles of comparable size (<4 nm) and 10 wt% loading, prepared by impregnation versus the double solvent methods and subsequent NaBH₄ reduction.²⁷⁴ Similar catalytic performance was observed for both syntheses with Ni/UiO-66, whereas impregnation was superior to the double solvent method for Ni encapsulated in MIL-101, in contradiction to the results of Zhen et al. The best performing Ni/MIL-101 catalyst prepared by impregnation gave 56% CO₂ conversion and 92% methane selectivity at 320 °C, a CO₂ : H₂ ratio of 8 and GHSV of 4650 h⁻¹. However, it is difficult to draw comparisons between these two studies due to their differing reaction conditions, Ni loading, particle sizes and synthetic details. Pd nanoparticles have also been dispersed in UiO-66 framework by a sol–gel method, and the resulting 6 wt% Pd/UiO-66 catalyst achieved 56% CO₂ conversion and 97% methane selectivity at 340 °C, 40 bar and a H₂ : CO₂ ratio of 4.²⁷⁷

Encapsulation of Ni nanoparticles in porous crystalline frameworks appears an effective strategy for CO₂ methanation, with 100% selectivity achieved over several catalysts, with highest yield (100% between 300–340 °C) reported for 20% Ni/MIL-101.¹¹¹ Deactivation mechanisms remain to be established, necessary to improve catalyst lifetimes for large-scale application. Synthetic methods able to maximise the fraction of Ni(111) facets in encapsulated nanoparticles should also be leveraged to promote low temperature CO₂ activation. To date, co-promoters have not been investigated for methanation catalysts derived from crystalline frameworks, and strategies to engineer e.g. hollow, core–shell, shape-controlled and bimetallic nanoparticles, or SACs have significant scope for exploitation in this field. The performance of porous crystalline framework derived methane selective catalysts prepared by various routes is compared in Table S3 (ESI†).

6.3 Carbon monoxide selective CO₂ reduction catalysts

The Reverse Water Gas Shift (RWGS) reaction, being endothermic, is favoured at high temperatures and hence is an energy intensive process. Catalysts are therefore sought for low temperature RWGS with high CO selectivity. Nanoparticle catalysts using noble metals such as Au, Pt and Rh show high activity for RWGS but offer poor CO selectivity at low temperature.²⁷⁸ In contrast, catalysts based on Earth-abundant metals such as Zn, Cu, Fe and In show poor activity and selectivity for the RWGS reaction.²⁷⁸ Attention has therefore focused on supported noble metal catalysts, with a recent shift to the use of crystalline frameworks as a matrix for their dispersion and activation.

Gutterod et al. investigated the performance of Pt nanoparticles encapsulated in a UiO-67 framework for RWGS. Although CO₂ conversion was only 6%, the selectivity to CO was >90% between 220 to 280 °C (along with a small amount of methane) for a H₂ : CO₂ ratio of 6 at 1 bar, however CO selectivity reached 100% under H₂ deficient conditions.²⁷⁹ For a loading of 1.5 wt%, ∼60% of Pt atoms were reduced at 350 °C, with a metal dispersion of 14% and 2.5 nm average particle size. Bimetallic Pt–Pd nanoparticles and Pt–Pd hollow nanocages encapsulated in UiO-67 were also studied, with Pt–Pd nanocages more active and selective, achieving 81% CO selectivity and 18% CO₂ conversion at 400 °C, 20 bar and a H₂ : CO₂ ratio of 3.²⁸⁰ CO was proposed to form by a redox mechanism, evidenced by a change in Pt²⁺/Pt⁰ speciation post-reaction. This performance was superior to Pt/mullite, Pt–Co/MCF-17 and Pt–Co/CeO₂ catalysts.²⁸⁰ Other bimetallic catalysts, comprising Au core–Pd shell (Au@Pd) nanoparticles encapsulated in different MOFs, namely UiO-66,²⁸¹ [Co₂(oba)₄(3-bpdh)₂]·4H₂O,²⁸² MOF-74²⁷⁸ and UiO-67²⁸⁰ have also been investigated for the RWGS reaction. The Au@Pd functionalised UiO-67 was also further functionalised with PVP-stabilised Pt nanoparticles. Catalysts were evaluated between 200 and 400 °C, at 20 bar and H₂ : CO₂ ratios ranging from 1 to 4. At 400 °C and a H₂ : CO₂ ratio of 4, CO₂ conversion decreased in the following order MOF-74 (35%) ∼ UiO-67 (35%) > Co-MOF (16%) > UiO-66 (7%), while CO selectivity decreased in the order MOF-74 (99%) > Co-MOF (88%) > UiO-67 (35%) > UiO-66 (72%). Pt nanoparticles on the external surface of the framework crystals were believed responsible for CO₂ activation while Au@Pd nanoparticles encapsulated within the MOFs conferred high CO selectivity. In all cases higher temperatures favoured methane production rather than CO. A redox mechanism for CO formation was proposed with synergy between the two types of metallic nanoparticles mediated by hydrogen spillover.

Noble metal nanoparticles, particularly bimetallics, when encapsulated in crystalline frameworks can effectively promote RWGS between 300–400 °C. Epitaxial growth of MOF shells around bimetallic nanoparticles, and promotion of the resulting materials by metal (Pt) nanoparticles, is a complex but creative design strategy to independently tune different steps of the reaction pathway. Judicious choice of an appropriate crystalline framework, such as MOF-74, facilitates CO₂ conversions reaching 35%, with CO selectivities exceeding 98%. Further investigations are required to elucidate the precise nature of the synergy between metal components within core–shell bimetallic, and between spatially separated metal nanoparticles, and how reactivity would be affected in the different metals were interchanged. A performance comparison of porous crystalline framework derived CO selective catalysts prepared by various routes appears in Table S4 (ESI†).

6.4 Formic acid selective CO₂ reduction catalysts

Formic acid (FA) is actively pursued as a renewable, liquid energy storage and transport medium due to its high effective gravimetric H₂ density (4.4 wt%) and low toxicity. FA decomposition can be controlled to achieve room temperature H₂ release over a suitable catalyst. Although metal complexes are reported for FA synthesis and decomposition, their practical application is hindered by their poor stability, requirement for solvents as a reaction medium, and problematic catalyst/product recovery and separation. Heterogeneous counterparts are therefore desirable. Pd mono-, bi- and trimetallic nanoparticles, such as PdAg, PdAu, PdCu, PdAuCo and PdCuCr, show high activity for CO₂ reduction to FA. Wen et al. reported CO₂ reduction to FA and subsequent FA decomposition using the same ZIF-8@Pd₁Ag₂@ZIF-8 catalyst in a liquid phase batch process.²⁸³ This appears the only example of thermocatalytic CO₂ reduction to FA using a catalyst derived from a crystalline framework. A “ship-in-a-bottle” approach was adopted to fabricate the catalyst; bimetallic Pd₁Ag₂ nanoparticles were added to solution during ZIF-8 synthesis, which was then allowed to proceed. This resulted in a layer of bimetallic Pd₁Ag₂ nanoparticles sandwiched between a ZIF-8 core and shell. The resulting FA productivity was 17 mmol_FA g_cat⁻¹ over 24 h at 100 °C, 20 bar and a H₂ : CO₂ ratio of 1. FA productivity decreased in the order ZIF-8@Pd₁Ag₂@ZIF-8 (17) > ZIF-8@Pd₂Ag₁@ZIF-8 (11) > ZIF-8@Pd_0.5Ag_2.5@ZIF-8 (10) > ZIF-8@Pd₃@ZIF-8 (8.5). The activity of the optimum Pd₁Ag₂ catalyst was attributed to the small size (2.8 nm) of bimetallic nanoparticles, enhanced CO₂ sorption in the ZIF-8 framework and charge-transfer to Pd from both Ag and proximate basic linkers. This catalyst was stable and reusable due to the protective ZIF-8 shell which hindered nanoparticles sintering.

6.5 Ethanol selective CO₂ reduction catalysts

Copper CO₂ reduction catalysts show poor selectivity towards ethanol and noble metals are required to enhance selectivity towards ethanol.²⁸⁴ An et al. designed an ethanol selective CO₂ reduction catalyst by supporting Cu^I sites on Zr₁₂ clusters, via deprotonation of hydroxyl groups on the clusters followed by metal–oxygen coordination, in a MOF assembled from Zr₁₂ SBUs and organic linkers. The 14 negative charges on the [Zr₁₂O₈(μ₃-O⁻)₈(μ₂-O⁻)₆(carboxylate)₁₈]¹⁴⁻ SBUs were balanced by 11 Cu^I species at SBU binding sites, and three neighbouring Li⁺ ions. The Li⁺ ions were subsequently ion-exchanged with Cs⁺ to obtain the active catalyst. The Zr₁₂-bpdc-CuCs catalyst was >99% selective to ethanol for 96% CO₂ conversion at 100 °C, 20 bar and a H₂ : CO₂ ratio of 3 in THF solvent. Activity rapidly dropped with increasing pressure, but under supercritical CO₂ (300 bar) a turnover number (TON) of 4080 was achieved in 10 hours (an order of magnitude higher than at 20 bar). Catalysis was ascribed to cooperativity between proximate dinuclear Cu^I₂ sites, which activated hydrogen and promoted C–C coupling of reactively-formed methanol and formyl species. Alkali cations provided an electron-rich environment for the Cu centre and helped stabilise the formyl intermediate. This study again highlights the importance of metal cluster-assembled crystalline frameworks as flexible architectures whose properties can be further refined to enable cooperative catalysis between well-defined active sites.

7.0 Crystalline frameworks as sacrificial precursors

Porous crystalline frameworks have also been explored as sacrificial precursors for the synthesis of catalysts with controlled composition, and unique nanostructure and morphology.^285–287 Pyrolysis (heating under an inert atmosphere) of frameworks is usually employed to obtain carbon derivatives, however annealing in the presence of sulphur, selenium or sodium hypophosphite has also been used to create sulphide, selenide and phosphide derivatives.²⁸⁵ The advantages of crystalline frameworks as sacrificial precursors are that: (i) they typically offer high surface areas, pore volumes and unique crystal morphologies, some of these properties being retained in their pyrolysed derivatives since decomposition of framework components occurs simultaneously throughout crystals; (ii) their composition, topology and morphology are tunable, which imparts a similar versatility in their derivatives; and (iii) guest molecules, nanoparticles and species incorporated in framework cavities may be retained in their derivatives. Core–shell morphologies are most commonly obtained by such approaches, wherein pyrolysis of organic linkers results in carbon encapsulation of node/guest species. This approach has also proved successful for fabricating carbon supported SACs.²²⁴

In the context of CO₂ reduction, framework-derived catalysts usually comprise porous functionalised carbon matrices encapsulating small (<5 nm) metal nanoparticles.²⁸⁸ Such carbon coated nanoparticles are resistant to sintering, phase change, and/or coking by virtue of the surrounding thin protective carbon layer. Chemical functionalities within the carbon shell can also influence nanoparticle reactivity,²⁸⁸ with CO₂ reduction to methanol, methane, CO and hydrocarbons all reported. A mind map of catalysts derived from crystalline frameworks for CO₂ reduction is presented in Fig. 15, with corresponding research reports compiled in Table S1 (ESI†).

Fig. 15 A mind map capturing representative catalysts derived from crystalline framework for CO₂ reduction to CH₃OH, CO, CH₄ and hydrocarbons.

7.1 Methanol selective CO₂ reduction catalysts

Decomposition of crystalline frameworks has been used to prepare carbon, metal oxide and alloy catalysts for the selective reduction of CO₂ to methanol. Zhang et al. synthesised a bimetallic CuO/ZnO catalyst dispersed in N-doped carbon by the pyrolysis of a Cu/Zn-BTC coordination polymer; this afforded a uniform dispersion of Cu and Zn in the pyrolysed catalyst and strong interfacial interactions.²⁸⁹ The catalytic activity was higher than a Cu/ZnO/Al₂O₃ reference for up to 275 hours on-stream, with comparable methanol selectivity obtained at 220–250 °C and 40 bar. The high catalyst stability was attributed to the high metal dispersion and Cu–Zn interactions. A Pd–ZnO catalyst derived from pyrolysis of ZIF-8 framework containing encapsulated Pd nanoparticles achieved a methanol productivity of 0.65 g_MeOH g_cat⁻¹ h⁻¹ at 270 °C, 45 bar and a H₂ : CO₂ ratio of 3.²⁹⁰ The encapsulated Pd nanoparticles spanned 2–4 nm, and evidence of PdZn alloy formation was apparent in the pyrolysed catalyst. Modest but stable CO₂ conversion (20%) and methanol selectivity (55%) were observed for 50 hours on-stream, in contrast to the parent Pd/ZIF-8 for which CO was the major product. The higher methanol selectivity observed for the MOF-derived Pd–ZnO catalyst likely reflects electronic perturbation of Pd upon alloying with Zn. Intermediate formate (HCOO*) species were detected over PdZn alloy particles, with ZnO species responsible for CO₂ activation and the stabilisation of oxygenate intermediates (H₂CO*, CH₃OH*).²⁹⁰

An In₂O₃@Co₃O₄ core–shell oxide composite, derived from the oxidative decomposition of In-impregnated ZIF-67(Co) achieved a similar methanol productivity of 0.65 g_MeOH g_cat⁻¹ h⁻¹ but with higher selectivity (87%) for 100 hours on-stream at 250–300 °C, 50 bar and a H₂ : CO₂ ratio of 4.¹¹⁵ During reaction, the core–shell oxide composite converted to a mixed carbide phase (Co₃InC_0.75) decorated by discrete cobalt and indium oxides. An induction period was observed for the onset of methanol production, associated with formation of the mixed carbide phase, although the oxide nanoparticles were considered the catalytically active species. The induction period could be altered by changing the synthesis and ZIF-67 decomposition protocol, decreasing its duration to only 10 hours (versus 30 hours for a co-precipitated catalyst). Methanol formation over In₂O₃ may follow a Mars–van Krevelen mechanism involving the generation of oxygen vacancies by H₂ and their quenching by CO₂ activation (promoted by Co₃O₄). A conventional In₂O₃/Co₃O₄ catalyst prepared by co-precipitation demonstrated 0.65 g_MeOH g_cat⁻¹ h⁻¹ and 80% selectivity, supporting the proposal that these oxides are the active species.

Liu et al. investigated a Cu@ZrO_x catalyst with 3D porosity, derived by in situ reconstruction of Cu nanoclusters encapsulated in defective UiO-66, obtaining a methanol productivity of 0.80 g_MeOH g_cat⁻¹ h⁻¹ at 13% CO₂ conversion and 79% selectivity, 260 °C, 45 bar and a H₂ : CO₂ ratio of 3, stable for 105 hours.²⁹¹ During reaction, defective Zr clusters of the UiO-66 framework rearranged into a stable tetragonal zirconia phase, while retaining strong Cu–ZrO_x metal support interactions through electron-transfer from Cu to Zr, which promoted the formation of Cu⁺ at the ZrO₂ interface. DRIFT studies suggest that Cu⁺ are responsible for the high methanol selectivity by virtue of stabilising reactively-formed formate (HCOO*) and methoxy (CH₃O*) intermediates.

ZrO₂–MO_x solid solutions comprising ZnO (ZnO–ZrO₂), CuO (CuO–ZrO₂) and Co₃O₄ (Co₃O₄–ZrO₂) obtained from metal-functionalised, Schiff base grafted UiO-66 precursors, were tested for CO₂ reduction at 320 °C, 30 bar and a H₂ : CO₂ ratio of 3 for 80 hours on-stream. The UiO-66 framework was post-synthetically modified to incorporate a Schiff base, with Zn, Cu or Co ions subsequently chelated to the base prior to pyrolysis to obtain the mixed oxide solid solutions. CuO–ZrO₂ showed the highest CO₂ conversion (23%) but was 87% selective to CO. DRIFTS showed the presence of a CO* intermediate on the catalyst surface, indicating poor stabilisation of desired oxygenates. Co₃O₄–ZrO₂ showed a modest CO₂ conversion of 11% but with 93% selectivity to methane, although only methoxy (CH₃O*) intermediates were observed by DRIFTS despite evident C–O bond cleavage. In contrast, ZnO–ZrO₂ showed low (6%) CO₂ conversion but 70% methanol selectivity, with formate and methoxy intermediates observed by DRIFTS; ZnO appears most effective in stabilising oxygenate intermediates.

Selective CO₂ reduction to methanol is possible over various mixed metal oxides derived from the decomposition of crystalline frameworks, including Pd–ZnO, In₂O₃–CO₃O₄–Co₃InC_0.75, Cu–ZrO_x and ZnO–ZrO₂ (though not CuO–ZrO_x). However, these catalysts appear to operate by diverse mechanisms involving alloys (Pd–ZnO), oxygen vacancy formation over redox active surfaces (In₂O₃–Co₃O₄) and selective stabilisation of methoxy intermediates (ZnO–ZrO_x). Although high methanol selectivity and good stability was demonstrated for some catalysts, comparable to conventionally prepared inorganic catalysts, improving their modest activity remains an ongoing challenge as evident from Table S2 provided in ESI.†

7.2 Methane selective CO₂ reduction catalysts

Porous crystalline frameworks have also been used as precursors to carbon and ZrO₂ catalysts selective for methane production. Lin et al. prepared a carbon catalyst containing a Ni nanoparticle core with a carbon shell, Ni@C, by pyrolysis of a Ni-MOF (of undisclosed structure but based on trimesic acid linkers) at 500 °C for 2 hours.²⁹² The resulting Ni@C catalyst was 100% selective to methane with a CO₂ conversion (25%) three times that of a conventional Ni/C catalyst at 250 °C. Higher reaction temperatures increased conversion to 92% with stable operation at 325 °C for 24 hours. The high methanation activity of Ru has been exploited through doping of UiO-66 and subsequent in-operando activation to form a 1 wt% Ru/ZrO₂ catalyst.²⁹³In situ XRD studies, supported by XAS, XPS and TEM, revealed that H₂ promotes the formation of Ru⁰ nanoparticles, while CO₂ acts as a mild oxidant to combust framework carbon and promote crystallisation of a (carbon-free) tetragonal ZrO₂ phase able to stabilise the highly dispersed Ru nanoparticles.²⁹⁴

MOF-derived Ru/ZrO₂ was completely selective to CH₄ at high H₂ (93%) and CO₂ (96%) conversions, and stable for 160 hours, outperforming a commercial methaniser catalyst and 1 wt% Ru/SBA-15, 1 wt% Ru/C and 1 wt% Ru/ZrO₂ controls under all reaction conditions. Zeng et al. investigated the performance of Ni supported on porous hydrous zirconia (obtained by UiO-66 decomposition in a basic medium) for CO₂ reduction in a fixed-bed reactor at 300 °C and 20 bar, achieving >99% selectivity and a methane productivity of 5851 mmol g_Ni⁻¹ h⁻¹.²⁹⁵ The parent framework morphology was retained in the porous ZrO₂, with surface hydroxyls reported to facilitate CO₂ activation, however no significant performance advantage was observed compared to conventional Ni catalysts. The performance of porous crystalline framework derived methane selective catalysts is compared in Table S3 (ESI†).

7.3 Carbon monoxide selective CO₂ reduction catalysts

Carbon catalysts derived from porous crystalline frameworks incorporating Cu, Zn and Co exhibit comparable RWGS performance to noble metal nanoparticles encapsulated within intact frameworks (Section 6.3). Zhang et al. investigated Cu/C and Cu/Zn/C catalysts prepared by pyrolysis of Cu-BTC frameworks as a strategy to hinder Cu nanoparticle sintering (held responsible for the deactivation of commercial Cu/Zn/Al₂O₃ catalysts).²⁹⁶ High metal (Cu and/or Zn) loadings of 70–75 wt% were achieved, albeit associated with large 15 to 25 nm particles. The pyrolysed carbon shell was porous, conferring surface areas spanning 96–125 m² g⁻¹ and protecting Cu nanoparticles against oxidation by air or N₂O. At 500 °C, a H₂ : CO₂ ratio of 3 and 1 bar, framework-derived Cu/C was 100% selective to CO with a stable, low activity for 20 hours. Catalyst stability was superior to a Cu/Zn catalyst prepared using an activated carbon support which deactivated after 11 hours on-stream. Lu et al. synthesized highly dispersed Co nanoparticles encapsulated in carbon matrices by pyrolysis of Co₃(BTC)₃·12H₂O or ZIF-67.²⁹⁷ The 700 °C pyrolyzed ZIF-67 derived catalyst (Co–C–N-700) contained Co nanoparticles in a N-doped carbon matrix, and was eight times more active (25% CO₂ conversion) for the RWGS reaction than a conventional Cu/Zn/Al₂O₃ catalyst, while achieving 90% CO selectivity at 450 °C, 1 bar and a H₂ : CO₂ ratio of 2. Co–C–N-700 also proved superior to a Co–C-700 analogue derived from Co₃(BTC)₃·12H₂O. The high activity of Co–C–N-700 catalyst was attributed to the formation of a Mott–Schottky junction and concomitant electron-transfer across the Co-support interface, which DFT calculations suggest lowers the barrier for CO₂ reduction to CO by a redox mechanism and formate intermediates.²⁹⁷ Although such reports are scarce, they highlight possible routes to noble metal-free catalysts for atmospheric pressure CO₂ reduction. A performance comparison of different porous crystalline framework derived CO selective catalysts prepared by various routes appears in Table S4 (ESI†).

7.4 CO₂ Fischer–Tropsch catalysts

CO₂ Fischer–Tropsch (CO₂-FT) synthesis has gained attention for CO₂ valorisation, with a Na-Fe₃O₄/HZSM-5 catalyst recently reported delivering up to 34% conversion and a high selectivity (up to 90%) to olefins and gasoline at 320 °C and 30 bar.³⁸ The use of crystalline frameworks as sacrificial precursors was initially reported for CO-FT synthesis to prepare catalysts based on iron and cobalt carbides encapsulated in a carbon matrix.^298–301 Guo et al. adopted a similar approach to disperse α-Fe₂O₃ nanoparticles in MIL-53(Al) and ZIF-8 by a solid grinding synthesis, resulting in a CO₂ conversion up to 30% to CO and higher hydrocarbons at 300 °C and 30 bar.³⁰² A higher olefin selectivity was observed for α-Fe₂O₃-ZIF-8 versus α-Fe₂O₃-MIL-53(Al). Decomposition of Fe-MIL-88b was also performed to obtain a high metal loading (34–51 wt%) CO₂-FT catalyst offering CO₂ conversion up to 46 wt% and good selectivity to hydrocarbons.³⁰³ Control over metal–support interactions is critical in CO₂ FT catalysis, since reduction and carbonation of must be balanced against the desire for a high dispersion (and hence active site density).²⁸⁸ The latter is aided by a classical SMSI which however negatively impacts on metal reducibility and CO₂ adsorption. Catalyst doping by K⁺ strongly promoted olefin formation relative to paraffins. During reaction, Fe nanoparticles were first reduced to a low valent/metallic state and subsequently a carbide phase whose composition was dependent on the nature of reaction intermediates and products.³⁰³ Pyrolysis at higher temperature resulted in metallic iron species which carbonized to a more graphitic, and less active, θ-Fe₃C phase during reaction, while pyrolysis at 500–600 °C predominantly generated magnetite and the in-operando formation of Haag Carbide (χ-Fe₅C₂) which was held responsible for FT activity. This carbide was favoured over the K-doped catalyst. Wezendonk et al. reached similar conclusions while investigating the impact of MOF topology on resulting Fe/C phases and their performance in CO-FT synthesis.³⁰¹ Fe/C catalysts derived by pyrolysis of MIL-68, MIL-88A, MIL-100, MIL-101, MIL-127, and Fe-BTC at 500 °C differed in their porosity, Fe content and nanoparticle size, crystallite density and thermal stability. Short nearest neighbour distances in parent frameworks led to larger Fe nanoparticles post pyrolysis, and a low pyrolysis temperature led to lower CO-FT activity. Iron catalysts (and K-promoted variants) derived from ZIF-8 for CO₂ reduction³⁰⁴ exhibited slightly lower conversion (36%) at 320 °C, 30 bar and a H₂ : CO₂ ratio of 3 than those obtained by MIL-88b pyrolysis,³⁰³ but achieved >60% hydrocarbon selectivity and a high olefin : paraffin ratio ≥5 due to the presence of pyridinic N.

Heteroatom promotion of iron carbide catalysts for CO₂-FT, derived by pyrolysis of Basolite F300 MOF, is also reported.²⁸⁸ A CO₂ conversion of 40% and selectivity to light olefins of 40% (24% selectivity to C₁₊ hydrocarbons) was obtained at 320 °C, 30 bar and a H₂ : CO ratio of 3 following promotion by 1 wt% K. Heteroatoms were introduced to the pyrolysed catalysts by incipient wetness impregnation. Despite the high (41 wt%) Fe loading, the average nanoparticle size was 4.4 nm with very few particles >20 nm. Potassium was the best promoter (being superior to Li and Na), although Cu and Pt also increased CO₂ conversion, CH₄ and light saturated hydrocarbons selectivity, whereas Mn, Mg and Mo enhanced CO selectivity and Ni, Rh and Co enhanced CH₄ selectivity. XRD and XPS showed the presence of sintered Fe₃O₄ and a complex mix of metallic iron, Fe₅C₂, Fe₇C₃ and Fe₃C in the K-promoted catalyst, whereas only less crystalline Fe₃O₄ and Fe₃C were observed in Pt and Mo promoted catalysts. CO₂-FT activity appears related to the RWGS activity of Fe₃O₄ and FT activity of Fe_xC_y phases. Potassium was largely present as the carbonate. CO₂ and H₂ chemisorption on the K-promoted catalyst were increased by 258% and decreased by 61% respectively compared to the unpromoted pyrolyzed catalyst. It was proposed that K charge-transfer to Fe strengthened the Fe–C interaction and rendered it less electrophilic, stabilising Fe against reduction by chemisorbed hydrogen. In contrast, Pt increased the H₂ uptake, likely by weakening Fe–C interactions and resulting in CH₄ formation due to a coverage of hydrogen adatoms. Mo-doping catalyst supressed CO₂ and H₂ adsorption, thereby promoting the RWGS reaction. In general, CO₂-FT generates a complex mixture of Fe metal, oxide and carbide phases, and greater control over iron speciation could result in further improvements in activity and selectivity to hydrocarbons. A performance comparison of porous crystalline framework derived CO₂ Fischer–Tropsch catalysts appears in Table S5 (ESI†).

8.0 Composite catalysts based on crystalline frameworks

Novel composites incorporating crystalline frameworks have appeared in the recent literature for niche applications, exploiting functional groups in the framework cavity or at their surface. Such composites may retain the physicochemical properties of their individual components, but gain additional properties such as liquid phase porosity,³⁰⁵ framework flexibility,³⁰⁶ and framework conductivity.³⁰⁷ Multifunctional and tandem reactions, including CO₂ reduction, can benefit from the presence of compositionally distinct entities. Only a few reports exist to date on composite catalysts derived from crystalline frameworks for CO₂ reduction, but significant advances are anticipated in this field.

8.1 Methanol selective CO₂ reduction catalysts

Zhao et al. investigated the performance of Cu–ZnO–Al₂O₃ catalysts prepared by the pyrolysis of a composite comprising a Cu/Al layered double hydroxide (stacked sheet morphology) and Zn-BTC MOF (nanoflower morphology). The resulting material was 90% selective to methanol at 200 °C, 30 bar and a H₂ : CO₂ ratio of 3, but only exhibited 4% CO₂ conversion.³⁰⁸ Conversion increased with reaction temperature, but at the expense of methanol selectivity. In situ DRIFTS identified HCOO*, CO* and CH₃O* intermediates, with CH₃O* stabilised over ZnO accounting for methanol formation. Pd/ZnO@ZIF-8 composites have also been prepared by encapsulation of Pd/ZnO microspheres³⁰⁹ or nanorods³¹⁰ with a ZIF-8 shell. CO₂ reduction over the nanorod composites achieved 80% methanol selectivity, albeit for a low (∼6%) conversion, at 250 °C, 45 bar and a H₂ : CO₂ ratio of 3 with stable activity over 25 hours. The ZIF-8 shell was grown by etching Zn²⁺ cations from the ZnO core to form the MOF precursor, which also introduced oxygen defects into the nanorods, with the ZIF-8 shell thickness controlled by the etching time. PdZn alloy nanoparticles were formed at the ZnO–ZIF-8 interface by subsequent wet impregnation and reduction. Methanol production was attributed to a synergy between the alloy nanoparticles and oxygen defects in ZnO. Increasing the ZIF-8 shell thickness created more defects, but disfavoured PdZn alloying at the expense of pure Pd nanoparticles. Metal-oxide@MOF core shell constructs are an attractive strategy for multistep reactions such as CO₂ reduction since active sites at the oxide core surface remain accessible through the porous (and potentially catalytically active) MOF shell.

8.2 Carbon monoxide selective CO₂ reduction catalysts

A Pd/ZSM-5@UiO-66-NH₂ composite catalyst, comprised of Pd/ZSM-5 and Pd/UiO-66-NH₂ is reported for RWGS at 320 °C, 20 bar and a H₂ : CO₂ ratio of 3 for 40 hours on-stream.³¹¹ CO₂ conversion decreased in the order Pd/UiO-66-NH₂ (25%) > Pd/ZSM-5@UiO-66-NH₂ (17%) > Pd/ZSM-5 (4%) > ZSM-5@UiO-66-NH₂ (0%), while CO selectivity decreased from Pd/ZSM-5@UiO-66-NH₂ (92%) > Pd/ZSM-5 (91%) > Pd/UiO-66-NH₂ (31%) > ZSM-5@UiO-66-NH₂ (0%). The Pd containing composite offered the optimal balance of modest conversion and high CO selectivity. The amine group in UiO-66-NH₂ helped to stabilise smaller Pd nanoparticles (∼2 nm) than could be obtained on ZSM-5 alone (5 nm); CO₂ conversion was proportional to Pd particle size. The high CO selectivity of Pd/ZSM-5 and Pd/ZSM-5@UiO-66-NH₂ compared to Pd/UiO-66-NH₂ indicated that ZSM-5 also participated in the reaction, possibly stabilising important reaction intermediates. CO₂ reduction was proposed to occur by a redox mechanism, evidenced by Pd²⁺ species on the catalyst post-reaction. Stabilisation of Pd nanoparticles by the UiO-66-NH₂ shell, and the high thermal stability of the ZSM-5 core, supressed catalyst deactivation, contrasting with Pd/UiO-66-NH₂ which decomposed to ZrO₂ during reaction. This study highlights the importance of crystalline framework linkers in stabilising metal nanoparticles.

8.3 Formic acid selective CO₂ reduction catalysts

Mori et al. investigated the performance of a PdAg/TiO₂@ZIF-8 composite catalyst for CO₂ reduction to formic acid in a batch reactor using 1 M NaHCO₃ as the CO₂ source at 100 °C, 20 bar and a H₂ : CO₂ ratio of 1 for 6 hours reaction, observing a TON of 913 and >99% formic acid selectivity.³¹² This TON was twice that of a standalone PdAg/TiO₂ catalyst, and 29 times higher than that of PdAg–ZIF-8. The nanocomposite was synthesised by encapsulating rutile crystallites, pre-functionalised with PdAg alloy nanoparticles, with a 1.6 nm thick ZIF-8 shell. DFT calculations suggest that alloy nanoparticles and Zn²⁺ ions in the ZIF-8 framework promote HCO₃⁻ adsorption, with the former also influencing the electronic density of carbon in the bound HCO₃⁻. Electronic-rich Pd was identified by XPS and DFT: the average atomic charge from DFT for Pd₁₉Ag₁₈ and Pd₁₁Ag₂₆ clusters in the presence of ZIF-8 was more negative that without the MOF shell. The high activity of PdAg/TiO₂@ZIF-8 versus PdAg/TiO₂ may in part reflect the presence of electronic-rich Pd in the composite, which reportedly promotes CO₂ reduction to formic acid.

9.0 Catalyst lifetime and deactivation

Catalyst lifetime is critical to the commercial implementation of a catalytic process, influencing overall sustainability, reactor selection, process design/integration, and of course economic viability.³¹³ Rapid catalyst deactivation and attendant short lifetimes can sometimes be overcome using recirculating fluidised bed reactors, coupled with a catalyst regeneration step, however this places additional constraints on catalyst formulation such as the need for binders, the choice of particle shape, mechanical strength, attrition resistance and so on. Catalysts derived from porous crystalline frameworks must therefore demonstrate sufficient stability and operational lifetime. Deactivation may arise from: (i) degradation of the metallic component due to sintering, coking, poisoning and/or oxidation;³¹⁴ (ii) physicochemical changes in the supporting or encapsulating crystalline framework.³¹⁵

9.1 Deactivation of metal component

Sintering is a general term used to describe the growth of metal particles under a reactive environment, and may occur through either particle migration and agglomeration or Ostwald ripening in which some particles grow at the expense of others.³¹⁶ It is a universal phenomenon in thermocatalysis processes, and while the extent may vary, it is almost always detrimental due to a loss of active surface area. Metal atom migration and/or particle agglomeration reflect the mobility of a material upon heating and are governed by the Tamman and Huttig temperatures.³¹⁷ Metal–support interactions can mitigate sintering, but may compromise catalytic activity and/or selectivity. For CO hydrogenation, monodispersed Ni nanoparticles over SiO₂ were transformed during reaction to a bimodal particle distribution comprising small spherical and large faceted crystals (with a high density of Ni(111) facets).³¹⁸ Particle sintering was attributed to the formation and migration of mobile Ni subcarbonyl species in the presence of CO.³¹⁸ The same phenomenon is potentially problematic for certain CO₂ reduction catalysts containing transition metals able to form volatile carbonyl complexes which facilitate support migration, or even loss to the vapor phase. In the case of Ni, improved nanoparticle stabilisation can be achieved using mixed oxides with a redox capacity, such as Ce_0.72Zr_0.28O₂,³¹⁹ and strongly interacting supports such as NiAl layered double hydroxides which allow for a topotactic transition of the Ni²⁺ ions,³²⁰ and can stabilise catalysts for hundreds of hour on-stream. Small (111) faceted Ni nanoparticles on MIL-101 also exhibit excellent stability for CO₂ reduction to methane for up to 200 hours at 300 °C.¹¹¹ Crystalline frameworks offer tunable, 3D framework–linker interactions for encapsulated metal particles, and hence better sinter-resistance than conventional low area, mesoporous or planar supports in which the metal local environment is less constrained.

Metal coking (surface poisoning by strongly-bound carbon) is another undesirable yet common mechanism of deactivation in thermal catalysts involving carbon-containing reactants,³²¹ and typically a consequence of cracking C–C and/or C–H bonds. Severe coking is typically observed at temperatures >300 °C, and its impact is especially severe for microporous catalysts, such as zeolites, and some MOFs and COFs, wherein small pores are rapidly blocked by carbon residues, and consequently any unique activity or e.g. shape selectivity, associated with micropores is lost.³²² Coking can be drastically reduced by low temperature operation, and or the introduction or pore hierarchy.^323,324 The high activity of catalysts derived from crystalline frameworks in some instances permits low(er) temperature CO₂ reduction, which may alleviate coking. The larger pore apertures and higher degree of pore connectivity in crystalline frameworks relative to zeolites is also expected to be beneficial in limiting the impact of coking, particularly for the CO₂-FT reaction wherein the production of higher carbons is often accompanied by coke formation. Iron-based CO₂-FT are also susceptible to phase changes on contact with carbonaceous species. The more stable, but less active, Fe₃C phase can form by carburisation of the more active, but less stable, Fe₅C₂ phase.³²² A better understanding of the stability and interconversion of iron carbide phases and their respective activities is important to advance CO₂-FT technologies.

Metal poisoning by reversibly-bound reactants, surface intermediates or products is difficult to control. Water and CO, common to almost all CO₂ reduction pathways, present a major challenge, reacting with transition and alkali/alkaline-earth metals. Water is a strong poison of Fe CO₂-FT catalysts due to its rapid reaction with Fe_xC phases, which suppresses hydrocarbon production.³²⁵ High pressure CO₂ feedstocks may also contain significant SO_x and NO_x impurities which can strongly coordinate metals within framework nodes and/or encapsulated or supported nanoparticles. Framework hydrophobisation is one strategy to improve the water tolerance of such catalysts. Further opportunities exist to construct multifunctional catalysts which combine the molecular sieving and selective binding capabilities of crystalline frameworks to remove SO_x and NO_x poisons, with catalytic sites for CO₂ reduction.

9.2 Reduction in activity due to change in physical and chemical characteristics of the crystalline framework

Different strategies have been proposed to utilise crystalline frameworks as catalyst components, where they serve the function of acting as a support, encapsulation materials or are present as a composite material for a specific purpose. Prolonged exposure to reaction environment can lead to changes in the physical and chemical characteristics of the framework components such as modification of framework–linker functional groups, modification of framework-nodes, partial destruction of framework structure due to breaking of weak bonds and permanent fixation of a rotating framework linker. Such changes may lead to temporary or permanent reduction in the activity and/or lifetime of crystalline frameworks-based catalysts. For example, a (Cu/ZnO)@MOF-5 catalyst completely collapsed during CO₂ reduction, with resulting methanol yields falling 17-fold (from 212 to 12 μmol_MeOH g_cat⁻¹ h⁻¹) due to this structural collapse.²⁷¹ A Ru/UiO-66 catalyst also collapsed during CO₂ reduction to form a Ru⁰/ZrO₂ phase, but retained 99.9% selectivity to methane and 96% CO₂ conversion for 160 hours.²⁹³ Such structural changes can impact the interaction of framework and metal components, potentially hindering the requisite synergy for cooperative catalysis in CO₂ reduction. For example, a study of condensation reactions in Fe(BTC) and MIL-100(Fe) MOFs revealed that reaction intermediates such as bisarylmethylcation, and by-products such as benzoic acid, can poison catalysis by strong adsorption and blockage of pores.³²⁶

10.0 Summary and outlook

The versatility of crystalline frameworks, which possess unique structural, physical and chemical properties has led to their rapid exploration as functional nanomaterials. CO₂ reduction presents a host of scientific and technical challenges, requiring multifunctional catalysts able to adsorb and activate CO₂ and H₂ at active sites within close proximity of each other, and which facilitate rapid molecular diffusion^327,328 and are stable under moderate temperatures and high pressures. The versatility of crystalline frameworks arises from the innumerable combinations of nodes and linkers (and ease of their chemical modification), and resulting diversity of cavity shape, size, topology and chemical composition.

Three promising design strategies have been exploited to develop CO₂ reduction catalysts derived from crystalline frameworks, namely the encapsulation (and/or dispersion) of metal, oxide or carbide nanoparticles, use of frameworks as sacrificial precursors, and the incorporation of frameworks into nanocomposites. The first of these may disrupt existing technologies for CO₂ reduction, evident from the high activity and tunable selectivity to either methanol, methane or CO. Encapsulation can increase the activity and stability of metal nanoparticles by SMSI and or metal–ligand charge-transfer with crystalline framework components. Confinement also ensures proximity between metal sites active for H₂ dissociation with those for CO₂ activation in framework cavities. For example, modulation of the electronic properties of Cu nanoparticles encapsulated in UiO-66 and UiO-66-bpy (Zr-based) frameworks,²³⁴ enabled methanol synthesis with high selectivity due to co-location of Zr/ZnO sites for CO₂ activation and Cu sites for H₂ activation. However, further improvements in activity (CO₂ conversion) are required for such framework-derived catalysts, while maintaining high methanol selectivity. This will necessitate precise control over the location and uniformity of encapsulated metal nanoparticles, and extended lifetime testing and synthetic scale-up. The exquisite structural control offered by porous crystalline frameworks is exemplified by the Zr₁₂-bpdc-CuCs catalyst for the selective reduction of CO₂ to ethanol,²⁸⁴ while CO₂ reduction to methane over a 20% Ni/MIL-101 catalyst (preferentially exposing Ni(111) facets) can deliver a 100% methane yield for 100 hours operation. The use of crystalline frameworks as sacrificial precursors to CO₂ reduction catalysts enables the preparation of metal core–porous carbon shell architectures that permit facile reactant/product transport to/from active sites while suppressing metal sintering and coking. Such materials are ideally suited for CO₂-FT to hydrocarbons (notably olefins).³⁰³ Controlled decomposition of crystalline frameworks has also allowed the synthesis of well-defined metal oxide catalysts with high activities for methanol¹¹⁵ or methane²⁹³ formation from CO₂. When carefully designed, porous crystalline frameworks can also be combined with other discrete catalyst entities such as layered double hydroxides³⁰⁸ or oxide nanocrystals³¹²e.g. or to form nanocomposite catalysts which retain the properties of their parent components but afford unique synergies, such as the promotion of CO₂ reduction to formic acid.

In light of the vast array of potential crystalline frameworks and functionalisation strategies, judicious selection of the appropriate CO₂ reduction catalyst should proceed through careful consideration of the precise reaction conditions (and hence product desired) under which it will operate. High temperatures and high H₂ : CO₂ ratios generally favour hydrocarbon products (CO₂-FT), but also increase the likelihood of coking. Low temperatures and low H₂ : CO₂ ratios (and the presence of basicity) will favour oxygenate products such as methanol and formic acid but result in poor activity.

As catalysts approach commercialisation, technoeconomic life cycle assessments will be necessary to assess the use of non-renewable H₂ for CO₂ reduction. New thermocatalytic technologies for CO₂ reduction may initially be exploited within the chemicals sector, which is responsible for significant CO₂ emissions (and wherein non-renewable H₂ production is already available) and is a major user of methanol, formic acid, hydrocarbons and syngas. CO₂ capture and use at such sources could drastically reduce the cost of feedstock/product supply and logistics. Technological progress in greenhouse gas capture and mitigation is likely to improve faster than solar, wind and battery advances.³²⁹ Such improvements will be driven in part by membrane separation, condensation/rectification and adsorption strategies.³²⁹ It remains to be seen whether scientific, engineering and regulatory agencies can facilitate adoption of such CO₂ capture strategies by emission intensive industries such as cement, iron, steel and ammonia production, petrochemical refining and fermentation. A supportive environment created by regulatory policies and appropriate carbon taxes or subsidies would promote the immediate adoption of thermocatalytic CO₂ reduction from non-renewable H₂ (grey H₂), offering partial mitigation of CO₂ emissions.

Advances in renewable H₂ (green H₂) production will significantly lower its price relative to existing grey H₂ sources (which are energy intensive with a large carbon footprint), driving the transition to cleaner CO₂ reduction processes using the former. A critical assessment of methanol production from CO₂ and green H₂ identified the cost of so-called green methanol to be 1.3 to 2.6 times higher than of fossil-derived methanol, with 73% of the green methanol costs associated with obtaining renewable H₂.³³⁰ Several innovative strategies have been proposed to improve the technoeconomic basis for green H₂.^331–333 For example, concentrated solar-thermal routes could yield H₂ at $3–17 per kg.^334,335 H₂ produced by water electrolysis using renewable electricity from wind, geothermal, tidal or hydro sources is generally considered the most sustainable and clean option, with costs between $2.91–10.92 per kg.⁵³ Current costs for grey H₂ are significantly lower ($1.35–3.5 per kg⁵⁴), hence the uptake of CO₂ reduction technologies is strongly dependent on lowering the cost of green H₂. However, the global landscape for green H₂ is set to drastically change in the near future. Expected strong reductions in capital expenditure costs for electrolysers of 70–80%, and falling levelized cost of renewable energy, are predicted to drive the cost of green H₂ to only $1–1.5 per kg in optimal locations (competitive with grey H₂) and $2–3 per kg under average conditions by 2030.⁵⁵ Worldwide governmental initiatives promoting large scale H₂ production will likely prove the biggest drivers to reduce the total cost of ownership for H₂ (production and distribution).⁵⁵ The development and implementation of large-scale thermocatalytic CO₂ reduction technologies, employing some form of crystalline framework catalysts, is therefore anticipated over the next two decades, when green H₂ is expected to be cost-competitive with alternative fossil and renewable energy resources.

A complete mechanistic understanding of thermocatalytic CO₂ reduction remains elusive. There is a wide gap between experimental studies and theoretical predictions. Although experiments have focused on composition–activity and structure–activity relationships, led by an empirical trial-and-error approach, few are sufficiently detailed and rigorous to inform theoretical chemists. Conversely, computational studies on simple model catalysts have limited application in rationalising experimental results from more complex nanoparticle and multifunctional catalysts in practical use. Therefore, it is vital for further progress that experimental and theoretical studies are conducted together in the same group or via collaborative efforts to further advance this exciting field.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

Authors sincerely acknowledge the support from Australian Research Council (ARC) under the Discovery project (DP180104076) on “Novel conversion process for carbon dioxide to chemicals”. R. B. acknowledges the Australian Research Council (DECRA fellowship DE160100987) for financial support.

References

1. J. R. Petit, J. Jouzel, D. Raynaud, N. I. Barkov, J. M. Barnola, I. Basile, M. Bender, J. Chappellaz, M. Davisk, G. Delaygue, M. Delmotte, V. M. Kotlyakov, M. Legrand, V. Y. Lipenkov, C. Lorius, L. P. Pin, C. Ritz, E. Saltzmank and M. Stievenard, Nature, 1999, 399, 429–436.
2. P. Lionello and L. Scarascia, Reg. Environ. Change, 2018, 18, 1481–1493.
3. R. Yu, P. Zhai and Y. Chen, Curr. Opin. Environ. Sust., 2018, 30, 75–81.
4. R. Falkner, Int. Aff., 2016, 92, 1107–1125.
5. S. Roy, A. Cherevotan and S. C. Peter, ACS Energy Lett., 2018, 3, 1938–1966.
6. M. D. Porosoff, B. Yan and J. G. Chen, Energy Environ. Sci., 2016, 9, 62–73.
7. M. Aresta, A. Dibenedetto and A. Angelini, Chem. Rev., 2014, 114, 1709–1742.
8. G. Pagani, US Pat., 4,208,347, 1980.
9. F. Zardi and S. Debernardi, US Pat., 8,367,867 B2, 2013.
10. H. Kolbe and E. Lautemann, Annalen, 1869, 113, 125.
11. E. Solvay, US Pat., 263981A, 1882.
12. M. W. Farlow and H. Adkins, J. Am. Chem. Soc., 1935, 57, 2222–2223.
13. B. Eliasson, C.-j. Liu and U. Kogelschatz, Ind. Eng. Chem. Res., 2000, 39, 1221–1227.
14. P. J. Chantry, J. Chem. Phys., 1972, 57, 3180–3186.
15. J. F. Walker and N. D. Scott, J. Am. Chem. Soc., 1938, 60, 951–955.
16. W. McGhee and D. Riley, J. Org. Chem., 1995, 60, 6205–6207.
17. Y. R. Jorapur and D. Y. Chi, J. Org. Chem., 2005, 70, 10774–10777.
18. M. A. Pacheco and C. L. Marshall, Energy Fuels, 1997, 11, 2–29.
19. T. Sakai, Y. Tsutsumi and T. Ema, Green Chem., 2008, 10, 337–341.
20. A. Rokicki and W. Kuran, J. Macromol. Sci., Rev. Macromol. Chem. Phys., 1981, 100, 135–186.
21. D. J. Darensbourg, Chem. Rev., 2007, 107, 2388–2410.
22. G.-P. Wu, S.-H. Wei, W.-M. Ren, X.-B. Lu, T.-Q. Xu and D. J. Darensbourg, J. Am. Chem. Soc., 2011, 133, 15191–15199.
23. M. Aresta, D. Ballivet-Tkatchenko, D. B. Dell'Amico, M. C. Bonnet, D. Boschi, F. Calderazzo, R. E. Faure, L. Labella and F. Marchetti, Chem. Commun., 2000, 1099–1100.
24. M. H. Jamróz, J. C. Dobrowolski, J. E. Rode and M. A. Borowiak, THEOCHEM, 2002, 618, 101–108.
25. A. Dibenedetto, M. Aresta, C. Fragale and M. Narracci, Green Chem., 2002, 4, 439–443.
26. O. Ihata, Y. Kayaki and T. Ikariya, Macromolecules, 2005, 38, 6429–6434.
27. I. Pápai, G. Schubert, I. Mayer, G. Besenyei and M. Aresta, Organometallics, 2004, 23, 5252–5259.
28. D. del Moral, A. M. Banet Osuna, A. Córdoba, J. M. Moretó, J. Veciana, S. Ricart and N. Ventosa, Chem. Commun., 2009, 4723–4725.
29. K. Buchmüller, N. Dahmen, E. Dinjus, D. Neumann, B. Powietzka, S. Pitter and J. Schön, Green Chem., 2003, 5, 218–223.
30. M. Aresta, A. Dibenedetto, I. Papai and G. Schubert, Inorg. Chim. Acta, 2002, 334, 294–300.
31. I. Yoshio, I. Yoshio and H. Harukichi, Chem. Lett., 1977, 855–856.
32. T. Tsuda, S. Morikawa, R. Sumiya and T. Saegusa, J. Org. Chem., 1988, 53, 3140–3145.
33. K. Armstrong and P. Styring, Front. Energy Res., 2015, 3, 8.
34. N. Mac Dowell, P. S. Fennell, N. Shah and G. C. Maitland, Nat. Clim. Change, 2017, 7, 243.
35. T. Witoon, P. Kidkhunthod, M. Chareonpanich and J. Limtrakul, Chem. Eng. J., 2018, 348, 713–722.
36. S. Bai, Q. Shao, P. Wang, Q. Dai, X. Wang and X. Huang, J. Am. Chem. Soc., 2017, 139, 6827–6830.
37. J. M. Thomas and K. D. M. Harris, Energy Environ. Sci., 2016, 9, 687–708.
38. J. Wei, Q. Ge, R. Yao, Z. Wen, C. Fang, L. Guo, H. Xu and J. Sun, Nat. Commun., 2017, 8, 15174.
39. C. Xie, C. Chen, Y. Yu, J. Su, Y. Li, G. A. Somorjai and P. Yang, Nano Lett., 2017, 17, 3798–3802.
40. Z. Li, J. Wang, Y. Qu, H. Liu, C. Tang, S. Miao, Z. Feng, H. An and C. Li, ACS Catal., 2017, 7, 8544–8548.
41. X. Zhang, A. Zhang, X. Jiang, J. Zhu, J. Liu, J. Li, G. Zhang, C. Song and X. Guo, J. CO₂ Util., 2019, 29, 140–145.
42. Y. Wang, L. Tan, M. Tan, P. Zhang, Y. Fang, Y. Yoneyama, G. Yang and N. Tsubaki, ACS Catal., 2019, 9(2), 895–901.
43. N. Thonemann and M. Pizzol, Energy Environ. Sci., 2019, 12, 2253–2263.
44. S. Deutz, D. Bongartz, B. Heuser, A. Kätelhön, L. Schulze Langenhorst, A. Omari, M. Walters, J. Klankermayer, W. Leitner, A. Mitsos, S. Pischinger and A. Bardow, Energy Environ. Sci., 2018, 11, 331–343.
45. R. G. Grim, Z. Huang, M. T. Guarnieri, J. R. Ferrell, L. Tao and J. A. Schaidle, Energy Environ. Sci., 2020, 13, 472–494.
46. M. Thema, F. Bauer and M. Sterner, Renewable Sustainable Energy Rev., 2019, 112, 775–787.
47. F. Schuth, Science, 2019, 363, 1282–1283.
48. T. Palden, B. Onghena, M. Regadío and K. Binnemans, Green Chem., 2019, 21, 5394–5404.
49. H. Rodhe, Science, 1990, 248, 1217–1219.
50. P. Bains, P. Psarras and J. Wilcox, Prog. Energy Combust. Sci., 2017, 63, 146–172.
51. E. I. Koytsoumpa, C. Bergins and E. Kakaras, J. Supercrit. Fluids, 2018, 132, 3–16.
52. C. Wang, G. Wang, R. Yang, X. Sun, H. Ma and S. Sun, Langmuir, 2017, 33, 503–509.
53. K. Schoots, F. Ferioli, G. Kramer and B. Vanderzwaan, Int. J. Hydrogen Energy, 2008, 33, 2630–2645.
54. H. Dagdougui, R. Sacile, C. Bersani and A. Ouammi, in Hydrogen Infrastructure for Energy Applications, ed. H. Dagdougui, R. Sacile, C. Bersani and A. Ouammi, Academic Press, 2018, pp. 7–21 DOI:10.1016/B978-0-12-812036-1.00002-0.
55. H. Council, Path to hydrogen competitiveness A cost perspective, 2020.
56. M. Ghoussoub, M. Xia, P. N. Duchesne, D. Segal and G. Ozin, Energy Environ. Sci., 2019, 12, 1122–1142.
57. Y. Song, X. Zhang, K. Xie, G. Wang and X. Bao, Adv. Mater., 2019, 31, 1902033.
58. L. Zhang, Z.-J. Zhao, T. Wang and J. Gong, Chem. Soc. Rev., 2018, 47, 5423–5443.
59. F. Marques Mota and D. H. Kim, Chem. Soc. Rev., 2019, 48, 205–259.
60. J. M. Spurgeon and B. Kumar, Energy Environ. Sci., 2018, 11, 1536–1551.
61. F. P. García de Arquer, C.-T. Dinh, A. Ozden, J. Wicks, C. McCallum, A. R. Kirmani, D.-H. Nam, C. Gabardo, A. Seifitokaldani, X. Wang, Y. C. Li, F. Li, J. Edwards, L. J. Richter, S. J. Thorpe, D. Sinton and E. H. Sargent, Science, 2020, 367, 661–666.
62. U. Ulmer, T. Dingle, P. N. Duchesne, R. H. Morris, A. Tavasoli, T. Wood and G. A. Ozin, Nat. Commun., 2019, 10, 3169.
63. Y. Ma, Z. Wang, X. Xu and J. Wang, Chin. J. Catal., 2017, 38, 1956–1969.
64. Y. Chen, D. Wang, X. Deng and Z. Li, Catal. Sci. Technol., 2017, 7, 4893–4904.
65. D. Chen, X. Zhang and A. F. Lee, J. Mater. Chem. A, 2015, 3, 14487–14516.
66. W. Wang, X. M. Xu, W. Zhou and Z. P. Shao, Adv. Sci., 2017, 4, 1600371.
67. A. Liu, M. Gao, X. Ren, F. Meng, Y. Yang, L. Gao, Q. Yang and T. Ma, J. Mater. Chem. A, 2020, 8, 3541–3562.
68. S. L. James, Chem. Soc. Rev., 2003, 32, 276–288.
69. X. Feng, X. Ding and D. Jiang, Chem. Soc. Rev., 2012, 41, 6010–6022.
70. J. Yao and H. Wang, Chem. Soc. Rev., 2014, 43, 4470–4493.
71. P. Z. Moghadam, A. Li, S. B. Wiggin, A. Tao, A. G. P. Maloney, P. A. Wood, S. C. Ward and D. Fairen-Jimenez, Chem. Mater., 2017, 29, 2618–2625.
72. J. L. Wang, C. Wang and W. B. Lin, ACS Catal., 2012, 2, 2630–2640.
73. S. B. Wang and X. C. Wang, Small, 2015, 11, 3097–3112.
74. A. Dhakshinamoorthy, A. M. Asiri and H. Garcia, Angew. Chem., Int. Ed., 2016, 55, 5414–5445.
75. L. Zeng, X. Y. Guo, C. He and C. Y. Duan, ACS Catal., 2016, 6, 7935–7947.
76. Z. B. Liang, C. Qu, D. G. Xia, R. Q. Zou and Q. Xu, Angew. Chem., Int. Ed., 2018, 57, 9604–9633.
77. J. D. Xiao, L. L. Han, J. Luo, S. H. Yu and H. L. Jiang, Angew. Chem., Int. Ed., 2018, 57, 1103–1107.
78. X. Chang, T. Wang and J. Gong, Energy Environ. Sci., 2016, 9, 2177–2196.
79. A. Mahmood, W. H. Guo, H. Tabassum and R. Q. Zou, Adv. Energy Mater., 2016, 6, 1600423.
80. B. A. Johnson, A. Bhunia, H. H. Fei, S. M. Cohen and S. Ott, J. Am. Chem. Soc., 2018, 140, 2985–2994.
81. D. Yang and B. C. Gates, ACS Catal., 2019, 9, 1779–1798.
82. L. Jiao, Y. Wang, H. L. Jiang and Q. Xu, Adv. Mater., 2018, 30, 1703663.
83. H. He, J. A. Perman, G. Zhu and S. Ma, Small, 2016, 12, 6309–6324.
84. Y. B. Huang, J. Liang, X. S. Wang and R. Cao, Chem. Soc. Rev., 2017, 46, 126–157.
85. A. Dhakshinamoorthy, A. M. Asiri and H. Garcia, ACS Catal., 2019, 9(2), 1081–1102.
86. Y.-S. Kang, Y. Lu, K. Chen, Y. Zhao, P. Wang and W.-Y. Sun, Coord. Chem. Rev., 2019, 378, 262–280.
87. J. Gascon, A. Corma, F. Kapteijn and F. X. Llabrés i Xamena, ACS Catal., 2014, 4, 361–378.
88. W.-G. Cui, G.-Y. Zhang, T.-L. Hu and X.-H. Bu, Coord. Chem. Rev., 2019, 387, 79–120.
89. C. Wang, B. An and W. Lin, ACS Catal., 2019, 9, 130–146.
90. F. M. Wisser, Y. Mohr, E. A. Quadrelli and J. Canivet, ChemCatChem, 2020, 12, 1270–1275.
91. M. Oschatz and M. Antonietti, Energy Environ. Sci., 2018, 11, 57–70.
92. Z. Bao, G. Chang, H. Xing, R. Krishna, Q. Ren and B. Chen, Energy Environ. Sci., 2016, 9, 3612–3641.
93. A. R. Millward and O. M. Yaghi, J. Am. Chem. Soc., 2005, 127, 17998–17999.
94. L. Xia, Q. Liu, F. Wang and J. Lu, J. Mol. Model., 2016, 22, 254.
95. A. Demessence, D. M. D'Alessandro, M. L. Foo and J. R. Long, J. Am. Chem. Soc., 2009, 131, 8784–8786.
96. C. Hon Lau, R. Babarao and M. R. Hill, Chem. Commun., 2013, 49, 3634–3636.
97. Y. Li, S. H. Chan and Q. Sun, Nanoscale, 2015, 7, 8663–8683.
98. R. Babarao and J. Jiang, Langmuir, 2008, 24, 6270–6278.
99. A. Stukowski, Modell. Simul. Mater. Sci. Eng., 2009, 18, 015012.
100. Y.-P. Chen, T.-F. Liu, S. Fordham and H.-C. Zhou, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2015, 71, 613–618.
101. A. Jain, S. P. Ong, G. Hautier, W. Chen, W. D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder and K. A. Persson, APL Mater., 2013, 1, 011002.
102. Y. A. Daza and J. N. Kuhn, RSC Adv., 2016, 6, 49675–49691.
103. X. Su, X. Yang, B. Zhao and Y. Huang, J. Energy Chem., 2017, 26, 854–867.
104. A. Álvarez, A. Bansode, A. Urakawa, A. V. Bavykina, T. A. Wezendonk, M. Makkee, J. Gascon and F. Kapteijn, Chem. Rev., 2017, 117, 9804–9838.
105. M. M.-J. Li and S. C. E. Tsang, Catal. Sci. Technol., 2018, 8, 3450–3464.
106. S. Dang, H. Yang, P. Gao, H. Wang, X. Li, W. Wei and Y. Sun, Catal. Today, 2019, 330, 61–75.
107. A. Dokania, A. Ramirez, A. Bavykina and J. Gascon, ACS Energy Lett., 2019, 4, 167–176.
108. S. Abelló, C. Berrueco and D. Montané, Fuel, 2013, 113, 598–609.
109. B. Lu and K. Kawamoto, Mater. Res. Bull., 2014, 53, 70.
110. K. Oshima, T. Shinagawa, Y. Nogami, R. Manabe, S. Ogo and Y. Sekine, Catal. Today, 2014, 232, 27–32.
111. W. Zhen, F. Gao, B. Tian, P. Ding, Y. Deng, Z. Li, H. Gao and G. Lu, J. Catal., 2017, 348, 200–211.
112. C. Janke, M. S. Duyar, M. Hoskins and R. Farrauto, Appl. Catal., B, 2014, 152–153, 184–191.
113. D. Sun, F. M. Khan and D. S. A. Simakov, Chem. Eng. J., 2017, 329, 165–177.
114. S. Rönsch, J. Schneider, S. Matthischke, M. Schlüter, M. Götz, J. Lefebvre, P. Prabhakaran and S. Bajohr, Fuel, 2016, 166, 276–296.
115. A. Pustovarenko, A. Dikhtiarenko, A. Bavykina, L. Gevers, A. Ramírez, A. Russkikh, S. Telalovic, A. Aguilar, J.-L. Hazemann, S. Ould-Chikh and J. Gascon, ACS Catal., 2020, 10, 5064–5076.
116. M. R. Rahimpour and M. Lotfinejad, Chem. Eng. Technol., 2007, 30, 1062–1076.
117. M. R. Rahimpour, M. Bayat and F. Rahmani, Chem. Eng. J., 2010, 157, 520–529.
118. J. Zhong, X. Yang, Z. Wu, B. Liang, Y. Huang and T. Zhang, Chem. Soc. Rev., 2020, 49, 1385–1413.
119. S. T. Sie and R. Krishna, Appl. Catal., A, 1999, 186, 55–70.
120. J.-S. Kim, S. Lee, S.-B. Lee, M.-J. Choi and K.-W. Lee, Catal. Today, 2006, 115, 228–234.
121. S. Ringe, E. L. Clark, J. Resasco, A. Walton, B. Seger, A. T. Bell and K. Chan, Energy Environ. Sci., 2019, 12, 3001–3014.
122. A. Bagger, W. Ju, A. S. Varela, P. Strasser and J. Rossmeisl, ACS Catal., 2019, 9, 7894–7899.
123. S. Xu and E. A. Carter, Chem. Rev., 2019, 119, 6631–6669.
124. S. S. Bhosale, A. K. Kharade, E. Jokar, A. Fathi, S.-m. Chang and E. W.-G. Diau, J. Am. Chem. Soc., 2019, 141, 20434–20442.
125. J. Fu, K. Jiang, X. Qiu, J. Yu and M. Liu, Mater. Today, 2020, 32, 222–243.
126. A. Alvarez, M. Borges, J. J. Corral-Perez, J. G. Olcina, L. Hu, D. Cornu, R. Huang, D. Stoian and A. Urakawa, ChemPhysChem, 2017, 18, 3135–3141.
127. X. Chen, X. Su, H.-Y. Su, X. Liu, S. Miao, Y. Zhao, K. Sun, Y. Huang and T. Zhang, ACS Catal., 2017, 7, 4613–4620.
128. J. Yoshihara and C. T. Campbell, J. Catal., 1996, 161, 776–782.
129. D. Cheng, F. R. Negreiros, E. Aprà and A. Fortunelli, ChemSusChem, 2013, 6, 944–965.
130. J. Yoshihara, S. C. Parker, A. Schafer and C. T. Campbell, Catal. Lett., 1995, 31, 313–324.
131. T. Kakumoto, Energy Convers. Manage., 1995, 36, 661–664.
132. Y. Yang, J. Evans, J. A. Rodriguez, M. G. White and P. Liu, Phys. Chem. Chem. Phys., 2010, 12, 9909–9917.
133. P. B. Rasmussen, P. M. Holmblad, T. Askgaard, C. V. Ovesen, P. Stoltze, J. K. Norskov and I. Chorkendorff, Catal. Lett., 1994, 26, 373–381.
134. I. A. Fisher and A. T. Bell, J. Catal., 1997, 172, 222–237.
135. T. Fujitani, I. Nakamura, T. Uchijima and J. Nakamura, Surf. Sci., 1997, 383, 285–298.
136. I. Nakamura, T. Fujitani, T. Uchijima and J. Nakamura, Surf. Sci., 1998, 400, 387–400.
137. S. Posada-Pérez, P. J. Ramírez, R. A. Gutiérrez, D. J. Stacchiola, F. Viñes, P. Liu, F. Illas and J. A. Rodriguez, Catal. Sci. Technol., 2016, 6, 6766–6777.
138. S. Collins, M. Baltanas and A. Bonivardi, J. Catal., 2004, 226, 410–421.
139. J. Ye, C.-j. Liu, D. Mei and Q. Ge, J. Catal., 2014, 317, 44–53.
140. S. Kattel, P. Liu and J. G. Chen, J. Am. Chem. Soc., 2017, 139, 9739–9754.
141. L. C. Grabow and M. Mavrikakis, ACS Catal., 2011, 1, 365–384.
142. M. Behrens, F. Studt, I. Kasatkin, S. Kuhl, M. Havecker, F. Abild-Pedersen, S. Zander, F. Girgsdies, P. Kurr, B.-L. Kniep, M. Tovar, R. W. Fischer, J. K. Norskov and R. Schlogl, Science, 2012, 336, 893–897.
143. Q.-L. Tang, Q.-J. Hong and Z.-P. Liu, J. Catal., 2009, 263, 114–122.
144. Q.-J. Hong and Z.-P. Liu, Surf. Sci., 2010, 604, 1869–1876.
145. Q. Sun and Z. Liu, Front. Chem., 2011, 6, 164.
146. S. Kattel, P. J. Ramirez, J. G. Chen, J. A. Rodriguez and P. Liu, Science, 2017, 355, 1296–1299.
147. C. Liu, B. Yang, E. Tyo, S. Seifert, J. DeBartolo, B. von Issendorff, P. Zapol, S. Vajda and L. A. Curtiss, J. Am. Chem. Soc., 2015, 137, 8676–8679.
148. J. Wang, S.-m. Lu, J. Li and C. Li, Chem. Commun., 2015, 51, 17615–17618.
149. X. Nie, X. Jiang, H. Wang, W. Luo, M. J. Janik, Y. Chen, X. Guo and C. Song, ACS Catal., 2018, 8, 4873–4892.
150. A. A. Gokhale, J. A. Dumesic and M. Mavrikakis, J. Am. Chem. Soc., 2008, 130, 1402–1414.
151. L. C. Grabow, A. A. Gokhale, S. T. Evans, J. A. Dumesic and M. Mavrikakis, J. Phys. Chem. C, 2008, 112, 4608–4617.
152. Y. Yang, M. G. White and P. Liu, J. Phys. Chem. C, 2012, 116, 248–256.
153. J. Graciani, K. Mudiyanselage, F. Xu, A. E. Baber, J. Evans, S. D. Senanayake, D. J. Stacchiola, P. Liu, J. Hrbek, J. F. Sanz and J. A. Rodriguez, Science, 2014, 345, 546–550.
154. S. Kattel, B. Yan, Y. Yang, J. G. Chen and P. Liu, J. Am. Chem. Soc., 2016, 138, 12440–12450.
155. R. Gaikwad, A. Bansode and A. Urakawa, J. Catal., 2016, 343, 127–132.
156. Y.-F. Zhao, Y. Yang, C. Mims, C. H. F. Peden, J. Li and D. Mei, J. Catal., 2011, 281, 199–211.
157. R. J. Madon, D. Braden, S. Kandoi, P. Nagel, M. Mavrikakis and J. A. Dumesic, J. Catal., 2011, 281, 1–11.
158. A. J. Elliott, R. A. Hadden, J. Tabatabaei, K. C. Waugh and F. W. Zemicael, J. Catal., 1995, 157, 153–161.
159. N. Pasupulety, H. Driss, Y. A. Alhamed, A. A. Alzahrani, M. A. Daous and L. Petrov, Appl. Catal., A, 2015, 504, 308–318.
160. Q. Tang, Z. Shen, L. Huang, T. He, H. Adidharma, A. G. Russell and M. Fan, Phys. Chem. Chem. Phys., 2017, 19, 18539–18555.
161. Y. Xia, Y. Xiong, B. Lim and S. E. Skrabalak, Angew. Chem., Int. Ed., 2009, 48, 60–103.
162. B. Anasori, M. R. Lukatskaya and Y. Gogotsi, Nat. Rev. Mater., 2017, 2, 16098.
163. X. W. Lou, L. A. Archer and Z. C. Yang, Adv. Mater., 2008, 20, 3987–4019.
164. S. C. Glotzer and M. J. Solomon, Nat. Mater., 2007, 6, 557–562.
165. C. M. A. Parlett, K. Wilson and A. F. Lee, Chem. Soc. Rev., 2013, 42, 3876–3893.
166. S. Zhou, L. Shang, Y. Zhao, R. Shi, G. I. N. Waterhouse, Y.-C. Huang, L. Zheng and T. Zhang, Adv. Mater., 2019, 0, 1900509.
167. S. F. J. Hackett, R. M. Brydson, M. H. Gass, I. Harvey, A. D. Newman, K. Wilson and A. F. Lee, Angew. Chem., Int. Ed., 2007, 46, 8593–8596.
168. K. Liu, X. Zhao, G. Ren, T. Yang, Y. Ren, A. F. Lee, Y. Su, X. Pan, J. Zhang, Z. Chen, J. Yang, X. Liu, T. Zhou, W. Xi, J. Luo, C. Zeng, H. Matsumoto, W. Liu, Q. Jiang, K. Wilson, A. Wang, B. Qiao, W. Li and T. Zhang, Nat. Commun., 2020, 11, 1263.
169. A. M. Abdel-Mageed, B. Rungtaweevoranit, M. Parlinska-Wojtan, X. Pei, O. M. Yaghi and R. J. Behm, J. Am. Chem. Soc., 2019, 141, 5201–5210.
170. K.-i. Otake, Y. Cui, C. T. Buru, Z. Li, J. T. Hupp and O. K. Farha, J. Am. Chem. Soc., 2018, 140, 8652–8656.
171. A. Nalaparaju, R. Babarao, X. S. Zhao and J. W. Jiang, ACS Nano, 2009, 3, 2563–2572.
172. W. Liang, R. Babarao, T. L. Church and D. M. D'Alessandro, Chem. Commun., 2015, 51, 11286–11289.
173. W. M. Bloch, R. Babarao, M. R. Hill, C. J. Doonan and C. J. Sumby, J. Am. Chem. Soc., 2013, 135, 10441–10448.
174. O. T. Qazvini, R. Babarao and S. G. Telfer, Chem. Mater., 2019, 31, 4919–4926.
175. L. K. Macreadie, E. J. Mensforth, R. Babarao, K. Konstas, S. G. Telfer, C. M. Doherty, J. Tsanaktsidis, S. R. Batten and M. R. Hill, J. Am. Chem. Soc., 2019, 141, 3828–3832.
176. V. E. Coyle, A. E. Kandjani, M. R. Field, P. Hartley, M. Chen, Y. M. Sabri and S. K. Bhargava, Biosens. Bioelectron., 2019, 141, 111479.
177. P. Makam, R. Shilpa, A. E. Kandjani, S. R. Periasamy, Y. M. Sabri, C. Madhu, S. K. Bhargava and T. Govindaraju, Biosens. Bioelectron., 2018, 100, 556–564.
178. A. Esmaielzadeh Kandjani, Y. M. Sabri, M. Mohammad-Taheri, V. Bansal and S. K. Bhargava, Environ. Sci. Technol., 2015, 49, 1578–1584.
179. A. E. Kandjani, M. Mohammadtaheri, A. Thakkar, S. K. Bhargava and V. Bansal, J. Colloid Interface Sci., 2014, 436, 251–257.
180. H. K. Daima, P. R. Selvakannan, A. E. Kandjani, R. Shukla, S. K. Bhargava and V. Bansal, Nanoscale, 2014, 6, 758–765.
181. Z. M. Davoudi, A. E. Kandjani, A. I. Bhatt, I. L. Kyratzis, A. P. O'Mullane and V. Bansal, Adv. Funct. Mater., 2014, 24, 1047–1053.
182. A. Elbourne, V. E. Coyle, V. K. Truong, Y. M. Sabri, A. E. Kandjani, S. K. Bhargava, E. P. Ivanova and R. J. Crawford, Nanoscale Adv., 2019, 1, 203–212.
183. V. Bansal, H. Jani, J. Du Plessis, P. J. Coloe and S. K. Bhargava, Adv. Mater., 2008, 20, 717–723.
184. S. Mehla, J. Das, D. Jampaiah, S. Periasamy, A. Nafady and S. K. Bhargava, Catal. Sci. Technol., 2019, 9, 3582–3602.
185. A. E. Kandjani, Y. M. Sabri, M. R. Field, V. E. Coyle, R. Smith and S. K. Bhargava, Chem. Mater., 2016, 28, 7919–7927.
186. Z. Hu, Y. Peng, Y. Gao, Y. Qian, S. Ying, D. Yuan, S. Horike, N. Ogiwara, R. Babarao, Y. Wang, N. Yan and D. Zhao, Chem. Mater., 2016, 28, 2659–2667.
187. S. Mehla, S. Kukade, P. Kumar, P. V. C. Rao, G. Sriganesh and R. Ravishankar, Fuel, 2019, 242, 487–495.
188. Y. Wang, H. Arandiyan, S. A. Bartlett, A. Trunschke, H. Sun, J. Scott, A. F. Lee, K. Wilson, T. Maschmeyer, R. Schlögl and R. Amal, Appl. Catal., B, 2020, 277, 119029.
189. M. D. Allendorf and V. Stavila, CrystEngComm, 2015, 17, 229–246.
190. W. Tu, Y. Xu, S. Yin and R. Xu, Adv. Mater., 2018, 30, 1707582.
191. O. K. Farha and J. T. Hupp, Acc. Chem. Res., 2010, 43, 1166–1175.
192. H. Furukawa, K. E. Cordova, M. O'Keeffe and O. M. Yaghi, Science, 2013, 341, 1230444.
193. S. Mehla, V. Krishsna, G. Sriganesh and R. Ravishankar, RSC Adv., 2018, 8, 33702–33709.
194. A. A. Olajire, Renewable Sustainable Energy Rev., 2018, 92, 570–607.
195. N. A. Khan, Z. Hasan and S. H. Jhung, Coord. Chem. Rev., 2018, 376, 20–45.
196. P. Silva, S. M. F. Vilela, J. P. C. Tome and F. A. A. Paz, Chem. Soc. Rev., 2015, 44, 6774–6803.
197. J. Chen, K. Shen and Y. Li, ChemSusChem, 2017, 10, 3165–3187.
198. C. Wang, D. M. Liu and W. B. Lin, J. Am. Chem. Soc., 2013, 135, 13222–13234.
199. R. Krishna, Chem. Soc. Rev., 2012, 41, 3099–3118.
200. W.-Y. Gao, A. D. Cardenal, C.-H. Wang and D. C. Powers, Chem. – Eur. J., 2019, 25, 3465–3476.
201. S. Lin, C. S. Diercks, Y.-B. Zhang, N. Kornienko, E. M. Nichols, Y. Zhao, A. R. Paris, D. Kim, P. Yang, O. M. Yaghi and C. J. Chang, Science, 2015, 349, 1208–1213.
202. H. Cui, Y. Wang, Y. Wang, Y.-Z. Fan, L. Zhang and C.-Y. Su, CrystEngComm, 2016, 18, 2203–2209.
203. Y. Lee, S. Kim, H. Fei, J. K. Kang and S. M. Cohen, Chem. Commun., 2015, 51, 16549–16552.
204. L. Shen, R. Liang, M. Luo, F. Jing and L. Wu, Phys. Chem. Chem. Phys., 2015, 17, 117–121.
205. H. J. Mackintosh, P. M. Budd and N. B. McKeown, J. Mater. Chem., 2008, 18, 573–578.
206. A. M. Shultz, O. K. Farha, J. T. Hupp and S. T. Nguyen, Chem. Sci., 2011, 2, 686–689.
207. X. Chen, M. Addicoat, E. Jin, L. Zhai, H. Xu, N. Huang, Z. Guo, L. Liu, S. Irle and D. Jiang, J. Am. Chem. Soc., 2015, 137, 3241–3247.
208. A. Nagai, X. Chen, X. Feng, X. Ding, Z. Guo and D. Jiang, Angew. Chem., Int. Ed., 2013, 52, 3770–3774.
209. W. Zhang, P. Jiang, Y. Wang, J. Zhang and P. Zhang, Catal. Sci. Technol., 2015, 5, 101–104.
210. E. D. Bloch, D. Britt, C. Lee, C. J. Doonan, F. J. Uribe-Romo, H. Furukawa, J. R. Long and O. M. Yaghi, J. Am. Chem. Soc., 2010, 132, 14382–14384.
211. J. X. Jiang, C. Wang, A. Laybourn, T. Hasell, R. Clowes, Y. Z. Khimyak, J. Xiao, S. J. Higgins, D. J. Adams and A. I. Cooper, Angew. Chem., Int. Ed., 2011, 50, 1072–1075.
212. W. Wang, A. Zheng, P. Zhao, C. Xia and F. Li, ACS Catal., 2014, 4, 321–327.
213. X. Han, Q. Xia, J. Huang, Y. Liu, C. Tan and Y. Cui, J. Am. Chem. Soc., 2017, 139, 8693–8697.
214. L.-H. Li, X.-L. Feng, X.-H. Cui, Y.-X. Ma, S.-Y. Ding and W. Wang, J. Am. Chem. Soc., 2017, 139, 6042–6045.
215. C. K. Brozek and M. Dincă, J. Am. Chem. Soc., 2013, 135, 12886–12891.
216. S. Yuan, Y.-P. Chen, J. Qin, W. Lu, X. Wang, Q. Zhang, M. Bosch, T.-F. Liu, X. Lian and H.-C. Zhou, Angew. Chem., Int. Ed., 2015, 54, 14696–14700.
217. Y. Ye, Z. Ma, R.-B. Lin, R. Krishna, W. Zhou, Q. Lin, Z. Zhang, S. Xiang and B. Chen, J. Am. Chem. Soc., 2019, 141, 4130–4136.
218. Y. Wang, X. Zhao, H. Yang, X. Bu, Y. Wang, X. Jia, J. Li and P. Feng, Angew. Chem., Int. Ed., 2019, 58, 6316.
219. J. C. Matsubu, V. N. Yang and P. Christopher, J. Am. Chem. Soc., 2015, 137, 3076–3084.
220. S. M. J. Rogge, A. Bavykina, J. Hajek, H. Garcia, A. I. Olivos-Suarez, A. Sepúlveda-Escribano, A. Vimont, G. Clet, P. Bazin, F. Kapteijn, M. Daturi, E. V. Ramos-Fernandez, F. X. Llabrés i Xamena, V. Van Speybroeck and J. Gascon, Chem. Soc. Rev., 2017, 46, 3134–3184.
221. Y.-S. Wei, M. Zhang, R. Zou and Q. Xu, Chem. Rev., 2020 DOI:10.1021/acs.chemrev.9b00757.
222. Ü. Kökçam-Demir, A. Goldman, L. Esrafili, M. Gharib, A. Morsali, O. Weingart and C. Janiak, Chem. Soc. Rev., 2020, 49, 2751–2798.
223. X. Li, X. Yang, Y. Huang, T. Zhang and B. Liu, Adv. Mater., 2019, 31, 1902031.
224. A. Han, B. Wang, A. Kumar, Y. Qin, J. Jin, X. Wang, C. Yang, B. Dong, Y. Jia, J. Liu and X. Sun, Small Methods, 2019, 3, 1800471.
225. M. F. M. Post, A. C. v. t. Hoog, J. K. Minderhoud and S. T. Sie, AIChE J., 1989, 35, 1107–1114.
226. D. Vervloet, F. Kapteijn, J. Nijenhuis and J. R. van Ommen, Catal. Sci. Technol., 2012, 2, 1221–1233.
227. Y. Yao, X. Liu, D. Hildebrandt and D. Glasser, Appl. Catal., A, 2012, 433–434, 58–68.
228. B. Wang, W. Liu, W. Zhang and J. Liu, Nano Res., 2017, 10, 3826–3835.
229. D. Liu, J. Wan, G. Pang and Z. Tang, Adv. Mater., 2019, 31, 1803291.
230. Z.-X. Cai, Z.-L. Wang, J. Kim and Y. Yamauchi, Adv. Mater., 2019, 31, 1804903.
231. N. Musselwhite, K. Na, K. Sabyrov, S. Alayoglu and G. A. Somorjai, J. Am. Chem. Soc., 2015, 137, 10231–10237.
232. A. Shivhare, J. A. Hunns, L. J. Durndell, C. M. A. Parlett, M. A. Isaacs, A. F. Lee and K. Wilson, ChemSusChem DOI:10.1002/cssc.202000764.
233. Y. Wang, S. Kattel, W. Gao, K. Li, P. Liu, J. G. Chen and H. Wang, Nat. Commun., 2019, 10, 1166.
234. B. An, J. Zhang, K. Cheng, P. Ji, C. Wang and W. Lin, J. Am. Chem. Soc., 2017, 139, 3834–3840.
235. S. Aguado, J. Canivet, Y. Schuurman and D. Farrusseng, J. Catal., 2011, 284, 207–214.
236. H. Li, C. Qiu, S. Ren, Q. Dong, S. Zhang, F. Zhou, X. Liang, J. Wang, S. Li and M. Yu, Science, 2020, 367, 667–671.
237. C. Yang, U. Kaipa, Q. Z. Mather, X. Wang, V. Nesterov, A. F. Venero and M. A. Omary, J. Am. Chem. Soc., 2011, 133, 18094–18097.
238. S. Yuan, L. Feng, K. C. Wang, J. D. Pang, M. Bosch, C. Lollar, Y. J. Sun, J. S. Qin, X. Y. Yang, P. Zhang, Q. Wang, L. F. Zou, Y. M. Zhang, L. L. Zhang, Y. Fang, J. L. Li and H. C. Zhou, Adv. Mater., 2018, 30, 1704303.
239. K. S. Park, Z. Ni, A. P. Cote, J. Y. Choi, R. Huang, F. J. Uribe-Romo, H. K. Chae, M. O'Keeffe and O. M. Yaghi, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 10186–10191.
240. N. M. Padial, E. Quartapelle Procopio, C. Montoro, E. Lopez, J. E. Oltra, V. Colombo, A. Maspero, N. Masciocchi, S. Galli, I. Senkovska, S. Kaskel, E. Barea and J. A. Navarro, Angew. Chem., Int. Ed., 2013, 52, 8290–8294.
241. B. Wang, X.-L. Lv, D. Feng, L.-H. Xie, J. Zhang, M. Li, Y. Xie, J.-R. Li and H.-C. Zhou, J. Am. Chem. Soc., 2016, 138, 6204–6216.
242. S. M. Cohen, Chem. Rev., 2012, 112, 970–1000.
243. S. M. Cohen, Chem. Sci., 2010, 1, 32–36.
244. H. Li, M. Eddaoudi, M. O'Keeffe and O. M. Yaghi, Nature, 1999, 402, 276–279.
245. M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. O'Keeffe and O. M. Yaghi, Science, 2002, 295, 469–472.
246. J. A. Greathouse and M. D. Allendorf, J. Am. Chem. Soc., 2006, 128, 10678–10679.
247. S. Zuluaga, E. M. A. Fuentes-Fernandez, K. Tan, F. Xu, J. Li, Y. J. Chabal and T. Thonhauser, J. Mater. Chem. A, 2016, 4, 5176–5183.
248. I. Pakamorė, J. Rousseau, C. Rousseau, E. Monflier and P. Á. Szilágyi, Green Chem., 2018, 20, 5292–5298.
249. M. Rubio-Martinez, C. Avci-Camur, A. W. Thornton, I. Imaz, D. Maspoch and M. R. Hill, Chem. Soc. Rev., 2017, 46, 3453–3480.
250. A. J. Howarth, Y. Liu, P. Li, Z. Li, T. C. Wang, J. T. Hupp and O. K. Farha, Nat. Rev. Mater., 2016, 1, 15018.
251. J. S. Qin, S. Yuan, C. Lollar, J. D. Pang, A. Alsalme and H. C. Zhou, Chem. Commun., 2018, 54, 4231–4249.
252. C. Serre, F. Millange, C. Thouvenot, M. Nogues, G. Marsolier, D. Louer and G. Ferey, J. Am. Chem. Soc., 2002, 124, 13519–13526.
253. G. Férey, C. Serre, C. Mellot-Draznieks, F. Millange, S. Surblé, J. Dutour and I. Margiolaki, Angew. Chem., 2004, 116, 6456–6461.
254. G. Ferey, C. Mellot-Draznieks, C. Serre, F. Millange, J. Dutour, S. Surble and I. Margiolaki, Science, 2005, 309, 2040–2042.
255. J. H. Cavka, S. Jakobsen, U. Olsbye, N. Guillou, C. Lamberti, S. Bordiga and K. P. Lillerud, J. Am. Chem. Soc., 2008, 130, 13850–13851.
256. D. Feng, Z.-Y. Gu, J.-R. Li, H.-L. Jiang, Z. Wei and H.-C. Zhou, Angew. Chem., Int. Ed., 2012, 51, 10307–10310.
257. W. Morris, B. Volosskiy, S. Demir, F. Gándara, P. L. McGrier, H. Furukawa, D. Cascio, J. F. Stoddart and O. M. Yaghi, Inorg. Chem., 2012, 51, 6443–6445.
258. K. Wang, X.-L. Lv, D. Feng, J. Li, S. Chen, J. Sun, L. Song, Y. Xie, J.-R. Li and H.-C. Zhou, J. Am. Chem. Soc., 2016, 138, 914–919.
259. H. J. Choi, M. Dinca and J. R. Long, J. Am. Chem. Soc., 2008, 130, 7848–7850.
260. X.-L. Lv, K. Wang, B. Wang, J. Su, X. Zou, Y. Xie, J.-R. Li and H.-C. Zhou, J. Am. Chem. Soc., 2017, 139, 211–217.
261. S. Wang, S. Hou, C. Wu, Y. Zhao and X. Ma, Chin. Chem. Lett., 2019, 30, 398.
262. C. Wu, F. Irshad, M. Luo, Y. Zhao, X. Ma and S. Wang, ChemCatChem, 2019, 11, 1256–1263.
263. C. D. Wu and M. Zhao, Adv. Mater., 2017, 29, 1605446.
264. M. Singh, D. Jampaiah, A. E. Kandjani, Y. M. Sabri, E. Della Gaspera, P. Reineck, M. Judd, J. Langley, N. Cox, J. van Embden, E. L. H. Mayes, B. C. Gibson, S. K. Bhargava, R. Ramanathan and V. Bansal, Nanoscale, 2018, 10, 6039–6050.
265. G. Li, S. Zhao, Y. Zhang and Z. Tang, Adv. Mater., 2018, 30, 1800702.
266. S. Luo, Z. Zeng, G. Zeng, Z. Liu, R. Xiao, M. Chen, L. Tang, W. Tang, C. Lai, M. Cheng, B. Shao, Q. Liang, H. Wang and D. Jiang, ACS Appl. Mater. Interfaces, 2019, 11, 32579–32598.
267. B. Rungtaweevoranit, J. Baek, J. R. Araujo, B. S. Archanjo, K. M. Choi, O. M. Yaghi and G. A. Somorjai, Nano Lett., 2016, 16, 7645–7649.
268. V. Mouarrawis, R. Plessius, J. I. van der Vlugt and J. N. H. Reek, Front. Chem., 2018, 6, 623.
269. P. J. Dauenhauer and O. A. Abdelrahman, ACS Cent. Sci., 2018, 4, 1235–1243.
270. T. L. Lohr and T. J. Marks, Nat. Chem., 2015, 7, 477–482.
271. M. Mueller, S. Hermes, K. Kaehler, M. W. E. van den Berg, M. Muhler and R. A. Fischer, Chem. Mater., 2008, 20, 4576–4587.
272. H. Tang, Y. Su, B. Zhang, A. F. Lee, M. A. Isaacs, K. Wilson, L. Li, Y. Ren, J. Huang, M. Haruta, B. Qiao, X. Liu, C. Jin, D. Su, J. Wang and T. Zhang, Sci. Adv., 2017, 3, e1700231.
273. H. Kobayashi, J. M. Taylor, Y. Mitsuka, N. Ogiwara, T. Yamamoto, T. Toriyama, S. Matsumura and H. Kitagawa, Chem. Sci., 2019, 10, 3289–3294.
274. M. Mihet, O. Grad, G. Blanita, T. Radu and M. D. Lazar, Int. J. Hydrogen Energy, 2019, 44, 13383–13396.
275. W. L. Zhen, B. Li, G. X. Lu and J. T. Ma, Chem. Commun., 2015, 51, 1728–1731.
276. Z.-W. Zhao, X. Zhou, Y.-N. Liu, C.-C. Shen, C.-Z. Yuan, Y.-F. Jiang, S.-J. Zhao, L.-B. Ma, T.-Y. Cheang and A.-W. Xu, Catal. Sci. Technol., 2018, 8, 3160–3165.
277. H. Jiang, Q. Gao, S. Wang, Y. Chen and M. Zhang, J. CO₂ Util., 2019, 31, 167–172.
278. Y. Han, H. Xu, Y. Su, Z.-l. Xu, K. Wang and W. Wang, J. Catal., 2019, 370, 70–78.
279. E. S. Gutterod, S. Oien-Odegaard, K. Bossers, A. E. Nieuwelink, M. Manzoli, L. Braglia, A. Lazzarini, E. Borfecchia, S. Ahmadigoltapeh, B. Bouchevreau, B. T. Lonstad-Bleken, R. Henry, C. Lamberti, S. Bordiga, B. M. Weckhuysen, K. P. Lillerud and U. Olsbye, Ind. Eng. Chem. Res., 2017, 56, 13207–13219.
280. H. Zhang, H. Xu, Y. Li and Y. Su, Appl. Mater. Today, 2020, 19, 100609.
281. Z. Zheng, H. Xu, Z. Xu and J. Ge, Small, 2018, 14, 1702812.
282. X. Zhao, H. Xu, X. Wang, Z. Zheng, Z. Xu and J. Ge, ACS Appl. Mater. Interfaces, 2018, 10, 15096–15103.
283. M. Wen, K. Mori, Y. Futamura, Y. Kuwahara, M. Navlani-García, T. An and H. Yamashita, Sci. Rep., 2019, 9, 15675.
284. B. An, Z. Li, Y. Song, J. Zhang, L. Zeng, C. Wang and W. Lin, Nat. Catal., 2019, 2, 709–717.
285. A. Indra, T. Song and U. Paik, Adv. Mater., 2018, 30, 1705146.
286. Q. Wang and D. Astruc, Chem. Rev., 2020, 120, 1438–1511.
287. K. O. Otun, X. Liu and D. Hildebrandt, J. Energy Chem., 2020, 51, 230–245.
288. A. Ramirez, L. Gevers, A. Bavykina, S. Ould-Chikh and J. Gascon, ACS Catal., 2018, 8, 9174–9182.
289. C. Zhang, P. Y. Liao, H. Wang, J. Sun and P. Gao, Mater. Chem. Phys., 2018, 215, 211–220.
290. Y. Yin, B. Hu, X. Li, X. Zhou, X. Hong and G. Liu, Appl. Catal., B, 2018, 234, 143–152.
291. T. Liu, X. Hong and G. Liu, ACS Catal., 2020, 10, 93–102.
292. X. Lin, S. Wang, W. Tu, Z. Hu, Z. Ding, Y. Hou, R. Xu and W. Dai, Catal. Sci. Technol., 2019, 9, 731–738.
293. R. Lippi, S. C. Howard, H. Barron, C. D. Easton, I. C. Madsen, L. J. Waddington, C. Vogt, M. R. Hill, C. J. Sumby, C. J. Doonan and D. F. Kennedy, J. Mater. Chem. A, 2017, 5, 12990–12997.
294. R. Lippi, A. M. D’Angelo, C. Li, S. C. Howard, I. C. Madsen, K. Wilson, A. F. Lee, C. J. Sumby, C. J. Doonan, J. Patel and D. F. Kennedy, Catal. Today, 2020 DOI:10.1016/j.cattod.2020.04.043.
295. L. Zeng, Y. Wang, Z. Li, Y. Song, J. Zhang, J. Wang, X. He, C. Wang and W. Lin, ACS Appl. Mater. Interfaces, 2020, 12, 17436–17442.
296. J. Zhang, B. An, Y. Hong, Y. Meng, X. Hu, C. Wang, J. Lin, W. Lin and Y. Wang, Mater. Chem. Front., 2017, 1, 2405–2409.
297. X. Lu, Y. Liu, Y. He, A. N. Kuhn, P.-C. Shih, C.-J. Sun, X. Wen, C. Shi and H. Yang, ACS Appl. Mater. Interfaces, 2019, 11, 27717–27726.
298. B. An, K. Cheng, C. Wang, Y. Wang and W. Lin, ACS Catal., 2016, 6, 3610–3618.
299. B. Qiu, C. Yang, W. Guo, Y. Xu, Z. Liang, D. Ma and R. Zou, J. Mater. Chem. A, 2017, 5, 8081–8086.
300. V. P. Santos, T. A. Wezendonk, J. J. D. Jaén, A. I. Dugulan, M. A. Nasalevich, H.-U. Islam, A. Chojecki, S. Sartipi, X. Sun, A. A. Hakeem, A. C. J. Koeken, M. Ruitenbeek, T. Davidian, G. R. Meima, G. Sankar, F. Kapteijn, M. Makkee and J. Gascon, Nat. Commun., 2015, 6, 6451.
301. T. A. Wezendonk, Q. S. E. Warringa, V. P. Santos, A. Chojecki, M. Ruitenbeek, G. Meima, M. Makkee, F. Kapteijn and J. Gascon, Faraday Discuss., 2017, 197, 225–242.
302. S. Hu, M. Liu, F. S. Ding, C. S. Song, G. L. Zhang and X. W. Guo, J. CO₂ Util., 2016, 15, 89–95.
303. J. H. Liu, A. F. Zhang, M. Liu, S. Hu, F. S. Ding, C. S. Song and X. W. Guo, J. CO₂ Util., 2017, 21, 100–107.
304. J. Liu, Y. Sun, X. Jiang, A. Zhang, C. Song and X. Guo, J. CO₂ Util., 2018, 25, 120–127.
305. M. Costa Gomes, L. Pison, C. Cervinka and A. Padua, Angew. Chem., Int. Ed., 2018, 57, 11909–11912.
306. M. L. Pinto, S. Dias and J. Pires, ACS Appl. Mater. Interfaces, 2013, 5, 2360–2363.
307. S. Chandra, T. Kundu, S. Kandambeth, R. BabaRao, Y. Marathe, S. M. Kunjir and R. Banerjee, J. Am. Chem. Soc., 2014, 136, 6570–6573.
308. F. Zhao, L. Fan, K. Xu, D. Hua, G. Zhan and S.-F. Zhou, J. CO₂ Util., 2019, 33, 222–232.
309. L. Lin, T. Zhang, H. Liu, J. Qiu and X. Zhang, Nanoscale, 2015, 7, 7615–7623.
310. X. Li, G. Liu, D. Xu, X. Hong and S. C. Edman Tsang, J. Mater. Chem. A, 2019, 7, 23878–23885.
311. X. Pan, H. Xu, X. Zhao and H. Zhang, ACS Sustainable Chem. Eng., 2020, 8, 1087–1094.
312. K. Mori, A. Konishi and H. Yamashita, J. Phys. Chem. C, 2020, 124(21), 11499–11505.
313. P. Forzatti and L. Lietti, Catal. Today, 1999, 52, 165–181.
314. A. Cao, R. Lu and G. Veser, Phys. Chem. Chem. Phys., 2010, 12, 13499–13510.
315. M. Opanasenko, A. Dhakshinamoorthy, J. Čejka and H. Garcia, ChemCatChem, 2013, 5, 1553–1561.
316. T. W. Hansen, A. T. DeLaRiva, S. R. Challa and A. K. Datye, Acc. Chem. Res., 2013, 46, 1720–1730.
317. A. J. Grano, F. M. Sayler, J.-H. Smått and M. G. Bakker, Mater. Lett., 2013, 111, 154–157.
318. M. Agnelli, M. Kolb and C. Mirodatos, J. Catal., 1994, 148, 9–21.
319. F. Ocampo, B. Louis and A.-C. Roger, Appl. Catal., A, 2009, 369, 90–96.
320. S. He, C. Li, H. Chen, D. Su, B. Zhang, X. Cao, B. Wang, M. Wei, D. G. Evans and X. Duan, Chem. Mater., 2013, 25, 1040–1046.
321. E. E. Wolf and F. Alfani, Catal. Rev., 1982, 24, 329–371.
322. S.-C. Lee, J.-S. Kim, W. C. Shin, M.-J. Choi and S.-J. Choung, J. Mol. Catal. A: Chem., 2009, 301, 98–105.
323. W. Khan, X. Jia, Z. Wu, J. Choi and A. C. Yip, Catalysts, 2019, 9, 127.
324. X. Shen, J. Kang, W. Niu, M. Wang, Q. Zhang and Y. Wang, Catal. Sci. Technol., 2017, 7, 3598–3612.
325. R. Satthawong, N. Koizumi, C. Song and P. Prasassarakich, Catal. Today, 2015, 251, 34–40.
326. M. Opanasenko, A. Dhakshinamoorthy, J. Čejka and H. Garcia, ChemCatChem, 2013, 5, 1553–1561.
327. D. Saha, Z. Bao, F. Jia and S. Deng, Environ. Sci. Technol., 2010, 44, 1820–1826.
328. T. M. Tovar, J. Zhao, W. T. Nunn, H. F. Barton, G. W. Peterson, G. N. Parsons and M. D. LeVan, J. Am. Chem. Soc., 2016, 138, 11449–11452.
329. M. Sharifzadeh, G. Triulzi and C. L. Magee, Energy Environ. Sci., 2019, 12, 2789–2805.
330. A. González-Garay, M. S. Frei, A. Al-Qahtani, C. Mondelli, G. Guillén-Gosálbez and J. Pérez-Ramírez, Energy Environ. Sci., 2019, 12, 3425–3436.
331. M. Qolipour, A. Mostafaeipour and O. M. Tousi, Renewable Sustainable Energy Rev., 2017, 78, 113–123.
332. R. Boudries, Int. J. Hydrogen Energy, 2018, 43, 3406–3417.
333. Y. Khojasteh Salkuyeh, B. A. Saville and H. L. MacLean, Int. J. Hydrogen Energy, 2017, 42, 18894–18909.
334. T. Pregger, D. Graf, W. Krewitt, C. Sattler, M. Roeb and S. Möller, Int. J. Hydrogen Energy, 2009, 34, 4256–4267.
335. R. J. Detz, J. N. H. Reek and B. C. C. van der Zwaan, Energy Environ. Sci., 2018, 11, 1653–1669.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ee01882a

---