Computer-aided bimetallic catalyst screening for ester selective hydrogenation

Heterogeneous hydrogenation of esters is a promising chemical process to produce alcohols. However, the selective hydrogenation of dibasic esters is still a challenge for both academia and industry. In this work, taking dimethyl oxalate (DMO) hydrogenation as an example, we have performed microkinetic analysis to explain the trend in the dimethyl oxalate hydrogenation activity and methyl glycolate (MG) selectivity across Ag, Cu, Ni, and Ru, using C and O adsorption energies as two descriptors. Ag is identified to be the best elemental metal catalyst for MG production. An unsupervised machine learning method based on the bisecting k -means hierarchical clustering algorithm is employed to determine the stable adsorption configurations over 1482 A 3 B 1 and 741 A 1 B 1 alloys. Ag 3 Zn 1 , Ag 3 Sn 1 , and Ag 3 Mg 1 catalysts are selected as promising bimetallic catalyst candidates due to their enhanced catalytic performance and relatively low cost. Bimetallic catalyst screening for ester selective hydrogenation has been performed by combining microkinetic analysis and machine learning methods.