A generalized machine learning framework to predict the space-time yield of methanol from thermocatalytic CO2 hydrogenation

Thermocatalytic CO2 hydrogenation to methanol is an attractive defossilization technology combat climate change while producing a valuable platform chemical and energy carrier. However, predicting the performance of catalytic systems for this process remains challenge. Herein, we present machine learning framework predict catalyst from experimental descriptors. A database Cu-, Pd-, In2O3-, ZnO-ZrO2-based catalysts with 1425 datapoints compiled literature subjected data mining. Accurate ensemble-tree models (R2 > 0.85) are developed space-time yield (STY) 12 descriptors, where significance space velocity, pressure, metal content revealed. The model prediction its insights experimentally validated, root mean squared error 0.11 gMeOH h?1 gcat?1 between actual predicted STY. purely data-driven, interpretable, cross-deployable other processes, serves as invaluable tool guided experiments optimization.