Key points are not available for this paper at this time.
Reaction mechanisms are at the core of understanding reaction systems and designing high-performance catalysts. A complex reaction system often involves various species and elementary reactions, posing a great challenge to determining the reaction mechanism. Here, we proposed a scheme to automatically generate reaction intermediates and elementary reactions to construct a complete reaction network represented by graph theory and employed a depth first search algorithm in the scheme to prune the reaction network to reduce the complexity of the network. With this scheme, microkinetic simulations of CO2 hydrogenation on Pd2Cu using the barriers predicted with the linear thermodynamics–kinetics relations were performed on the network to determine the mechanism and rate- and selectivity-controlling steps of CO2 hydrogenation to ethanol and methanol. Analysis shows that the simulated selectivity of ethanol and methanol agrees well with the experimental results. CO2 + H → COOH is the rate-controlling step, and CHOH + H → CH + H2O, CH2OH + H → CH2 + H2O, and CH2OH + H → CH3OH dominate the ethanol selectivity. Both ethanol and methanol are generated via multiple reaction pathway mechanisms. Investigations of the pruned networks show that quantitatively correct results can be obtained from the pruned or pseudocomplete reaction network, as long as the key pathways are embodied in the network. 94% ethanol selectivity of the complete network can be obtained with the pruned network composed of 60 elementary steps, compared to 176 steps of the complete network. The present work articulates graph theory representation, depth first search algorithm, linear thermodynamics–kinetics relations, and microkinetic simulations to approach complicated heterogeneous reaction systems and exemplifies their comprehensive roles in exploring complex reaction networks.
Guo et al. (Mon,) studied this question.