Abstract
The excellent performance of ZnxCryOz catalysts, used in the process of converting CO/H2 to methanol and light olefins, is highly attractive, but the microstructure of ZnxCryOz structure under the syngas conversion conditions remains elusive experimentally and theoretically because of the limitation of the detecting facilities/methods. By using the genetic-algorithm-based global structural search accelerated by machine learning in combination with a local cluster sampling strategy in the active learning scheme, we reveal the structure/composition evolution of ZnxCryOz structures and uncover that the catalytic activities of these catalysts strongly depend on the Zn/Cr ratios under the syngas conversion conditions. The possible active phase at the thermodynamically stable condition is identified and the critical active site influencing the catalytic property is unraveled. We show that the catalyst Zn2Cr2O5, which consists of a thin ZnO layer and the ZnCr2O4 structures, achieves a high catalytic activity for syngas conversion and its X-ray diffraction patterns are in agreement with the experimental result. Importantly, the presence of a hexahedral configuration ([ZnCrO2]hex) is found to affect the catalyst activity significantly, and this result is further supported by the analysis based on the structure-activity relationship.
Supplementary materials
Title
Identifying the Active Site on ZnxCryOz for HC-O Bond Cleavage in Syn-gas Conversion
Description
Supporting Information
Contents
1. Reduced ZnxCryOz surfaces with the surface Zn/Cr=1/2~2
2. Surface properties of Zn2Cr2O5 structures
3. ZnO doping structures
4. Neural network (NN) potential
Actions