Experimental data-driven reaction network identification and uncertainty quantification of CO2-assisted ethane dehydrogenation over Ga2O3/Al2O3

Weiqi Chen,Maximilian Cohen,Kewei Yu,Hsuan-Lan Wang,Weiqing Zheng,Dionisios G. Vlachos

DOI: https://doi.org/10.1016/j.ces.2021.116534

IF: 4.7

2021-06-01

Chemical Engineering Science

Abstract:We study ethane dehydrogenation's kinetics over a Ga2O3/Al2O3 catalyst in the absence and presence of CO2 and H2O. We identify the reaction network through the hypothesis of overall-reactions and reaction stoichiometry using judicious experiments by co-feeding combinations of C2H6, CO2, H2, and H2O and regeneration studies. We introduce an uncertainty quantification methodology, leveraging Bayesian inference, to assess the rate parameters' statistical confidence and trace the uncertainty sources and develop a data-driven kinetics model. The chief overall reactions include ethane dehydrogenation and hydrogenolysis, the reverse water–gas shift (RWGS), coke formation, and coke gasification. H2O retards the ethane dehydrogenation rate and increases the apparent activation energy from 52.4 to 177.3 kJ/mol. Both CO2 and H2O improve catalyst stability via the same mechanism: coke gasification. Raman characterization of the spent catalyst reveals nanocrystal graphite coke whose amount is reduced by H2O. The reaction-network identification and uncertainty quantification apply to other complex reaction systems.

engineering, chemical