Generative Design of Oxides as CO2 Hydrogenation Catalysts

A key goal in sustainable development is to find ways to convert carbon dioxide (CO2) into valuable chemicals. One critical reaction is the conversion of CO2 with H2 to produce methanol, which can be used as a fuel or precursor of various important chemicals. This process is enabled by materials called catalysts, which accelerate the rate of chemical reactions and increase their efficiency toward producing the desired products. Binary oxides comprising two metals and oxygen have emerged as a promising alternative to standard metal-based catalysts for forming methanol from CO2 and H2. However, their complexity makes it challenging to fully understand how their composition and structure relate to their catalytic activity. Today, computer simulations are commonly used to understand how catalysts work at the atomic level. Still, they are computationally intensive and costly, limiting the understanding and the exploration of alternative materials. The GENOX project aims to identify the critical properties of mixed metal oxides to design and predict better materials for hydrogenating CO2 to methanol. Several catalytic materials will be evaluated using data-driven and artificial intelligence (AI) methods to establish connections between computed and experimental properties and their catalytic activity. After validating the accuracy of the methods by benchmarking against experimental data from collaborators who are experts in the field, we will use generative AI to suggest better catalytic materials. Overall, the GENOX project takes an alternative approach driven by artificial intelligence to unravel the properties of oxide materials driving their catalytic activity for the sustainable conversion of CO2.