上海科技大学知识管理系统

Reactions on the surface of catalysts are rather complex, and many possible reaction pathways and intermediates are involved. We here propose a method that is able to automatically generate a catalytic reaction network and identify the preferred reaction pathway with determined uncertainty. Taking syngas conversion to ethanol on Rh(111) as an example, a reaction network consisting of 95 elementary steps was generated. Using energies calculated with an ensemble of 2000 functionals, the occurrence frequency of different ethanol formation pathways was obtained through pruning of the reaction network with mean-field microkinetic modeling. We found that CHCO is the most important reaction intermediate for ethanol formation with the highest confidence, even at varied temperatures and pressures. The transition state of CH3CH2O hydrogenation, i.e. CH3CH2O-H, possesses the highest possibility to be rate-controlling. CO has the highest possibility to be the surface dominant species at all the temperatures and pressures considered. The method developed in the current work substantially reduces the complexity of identifying the mechanism of catalytic reactions and shows great potential in expediting future catalyst design.