Identifying Catalyst Property Descriptors for CO2 Hydrogenation to Methanol via Big-Data Analysis

Carbon dioxide (CO2) capture and valorization have great potential for mitigating emissions of this greenhouse gas and accordingly for preserving the environment for future generations. In this regard, hydrogenation of CO2 to methanol is highly attractive because this product is a valuable energy carrier and can also be used for production of various everyday commodities. Although many research papers on this topic have been published in the past decades, there is still a lack of fundamentals relevant to control catalyst performance. Herein, we demonstrate how statistically validated Big-Data analysis of available literature data identified hidden descriptors that can be applied for purposeful catalyst development and for identification of optimal reaction conditions. In view of catalyst development, the kinds of structural promoters or supports for bulk or supported Cu-, In-, or Pd-based catalysts are the most important descriptors for methanol selectivity, with Ce and Zr being the most efficient promoters. The type and the parameters of the preparation methods as well as the kind of active component precursors are also important in this regard. To validate the conclusion about the structural promoter, a series of supported CuZn-containing catalysts were prepared. The best-performing CuZn/CeO2 catalyst outperformed the state-of-the-art CuZn-based catalysts tested at a total pressure of up to 30 bar using a feed with the ratio of H2/CO2 of 3. In addition to the catalyst composition and the preparation method, our analysis suggests that the most often used Cu-based catalysts lose their methanol selectivity due to the decomposition of this product to CO. Our control experiments with the developed CuZn-based catalysts proved that this undesired reaction can be hindered when the catalyst support contains Ce or through increasing H2 partial pressure. This knowledge is important for further catalyst development.