Rational catalyst design for transforming CO2 into industrially attractive products: Formic acid, polycarbonates and polyurethanes.

The dramatic impact of massive carbon dioxide (CO2) emissions on climate change prompts the development of new energy sources and recycling strategies. This project has focused on the latter approach by exploring the use of this greenhouse gas in the production of chemical feedstocks and materials with added value. These include the hydrogenation of CO2 to formic acid and methanol by metal-ligand bifunctional catalysts and by functionalized metal organic frameworks (MOFs). The mechanisms of these reactions have been studied by using quantum mechanics calculations based on density functional theory (DFT). The hydrogenation of CO2 to methanol catalyzed by pincer Fe complexes and amines involves the formation of amides. In order to optimize this reaction we studied the mechanism for the hydrogenation of amides in collaboration with the experimental groups of Nilay Hazari (Yale University, USA) and Wesley Bernskoetter (University of Missouri, USA). Our results suggested that the mechanism is amide-dependent (primary and secondary amides reacts by different pathways) and primary amides co-catalyze the H-transfer and C-N bond cleavage required for this transformation (ACS Catal., 2018, 8, 8751). A drawback of using primary amides as co-catalyst is that they are consumed during the reaction and can poison the catalyst. Considering this information, we explored computationally other substances that could act as co-catalyst but preventing these side reactions (Chem. Sci., 2020, 11, 2225). The co-catalysts leading to the best computational results were tested experimentally and improved the performance of the reaction. In order to get insight on the influence of the metal in the hydrogenation of amides, a Mo(II) catalyst was studied in collaboration with the group of Matthias Beller (LIKAT, Germany). The DFT study showed that the active Mo(0) species can reduce the C=O group of amides through lower energy barriers than the Fe-based system (Chem. Sci., 2019, 10, 10566). However, the alcohol produced poison the catalyst resulting in a less efficient system. These results indicate that further catalyst design should focus on preventing the formation of adducts, while keeping the high hydricity of the complex. In collaboration with the catalysis group at the University of Oslo (UiO), we evaluated the possibility of supporting homogenous catalysts able to hydrogenate CO2 to methanol in MOFs, to make these catalysts industrially attractive. The synthesis of these materials is however highly challenging. Our first approach was to incorporate Mg atoms in the Zr-nodes of UiO-66 and -67 MOFs. The computational study of these systems revealed that replacing the m-OH groups by ?-OMgCH3 changes significantly the energy and volume of the resulting three configurations, which contain the m-OMgCH3 groups in different relative positions. This result is relevant for understanding the reactivity of node-functionalized MOF materials, which are used in catalytic transformations. The incorporation of amines to these systems was predicted to promote the activation of CO2. However, the poor crystallinity of this material made difficult its further study and characterization. In addition to holding molecular catalysts, MOFs can be used to encapsulate metal nano-particles (NPs) and change its reactivity. This behavior was observed when introducing Pt-NPs in UiO-67 MOF, which allowed the hydrogenation of CO2 to methanol (J. Am. Chem. Soc. 2020, 142, 999). Steady-state and transient kinetic studies coupled with operando infrared spectroscopy showed that Zr-nodes play a key role in the formation of methanol. DFT calculations showed that linkers are easily displaced in close contact with Pt-NPs giving access to the Zr-active sites. The combination of experiments and calculations also gave insight into the influence of water in the reaction (J. Am. Chem. Soc. 2020, 142, 17105). In this case, a microkinetic model on the methanol and water competition for the Zr-active sites, based on calculated energies, helped to rationalize the experimental observations. Thus, the higher formation of methanol in dehydrated and highly defectives nodes could be explained by the weaker adsorption of methanol in these systems; and the non-deuteration of MeOH using D2O was consistent with the computed MeOH desorption mechanism. These results will be useful for the design of new MOF-based catalysts for CO2 hydrogenation reactions. In conclusion, over five years, the CaRD project has contributed to the design of more efficient homogeneous and heterogeneous catalysts for the conversion of CO2 to methanol and in the mechanistic understanding of this complex reaction.

This project has shown the importance of using microkinetic models for the study of hydrogenation reactions that imply several catalytic cycles. We expect that after our results this methodology becomes more frequent to study hydrogenation reactions involving homogeneous and MOF-based catalysts. The publications produced by this project has turned the University of Oslo (UiO) into a good partner for projects involving CO2 conversion. After CaRD, the project leader has been involved in several applications on CO2 reductions projects, and succeed in the creation of the "Nordic consortium for CO2 conversion" (NordCO2) and the innovative training network "Cooperation towards a sustainable chemistry industry" (CO2PERATE). These networks will impact in the education received by several students in CO2 conversion, increasing interdisciplinary research and involving international collaboration, and making these students highly competent for working in both industry and academia.

Using CO2 as a building block in organic synthesis will assist in the development of a renewable carbon economy and is potentially a sustainable strategy for CO2 capture and storage (CCS). In this project, two reactions with a high percentage of CO2 in the product will be explored in detail: the hydrogenation of CO2 to form formic acid and the generation of poly(cyclohexene-alt-carbonate) from CO2 and epoxides. These reactions are catalyzed by Fe and Ti compounds supported by tridentate ligands with PNP and OCO donors, respectively. However, efficiency, selectivity and robustness need to be improved in order to make them industrially attractive. Strategies to address these problems will be developed using computational chemistry. The catalysts will be studied by determining the reaction mechanisms of both productive and deactivation pathways. This knowledge will be used to further optimize these catalysts and to design a new efficient catalyst for a third unprecedented reaction, the formation of polycarbamates from CO2 and imines. The success of this reaction will give access to a new class of polyurethanes, of unknown properties, by a green process mediated by CO2. This reaction will be constructed by considering the elementary steps of the catalytic cycle computationally and experimentally.