Kinetic Monte Carlo simulations of heterogeneously catalyzed oxidation reactions

Abstract

In this perspective, we focus on the catalyzed oxidation of CO and HCl over the model catalyst RuO2(110) and how the kinetics of these reactions can only properly be modeled by kinetic Monte Carlo (kMC) simulations when lateral interactions of the surface species are taken into account.

Combined experiment and theory approach in surface chemistry: Stairway to heaven?

“…However, in the transient region starting from a stoichiometric RuO 2 (110) surface and exposing the surface to p(CO) = 2 × 10 −7 mbar and p(O 2 ) = 10 −7 mbar both parameter sets result in distinctly different surface configurations, as shown in Fig. 6c and in movies provided on our website [85,86]. In Seitsonen's parameter set, separated CO-and O-covered islands are formed, where the recombination of O and CO takes place predominantly in the lowcoverage boundary regions between the CO and O islands.…”

Kinetic models of CO oxidation on gold nanoparticles

“…In this context, we may notice that the analytical treatments (like those in our analysis) of the kinetics of reactions occurring on a small number of different catalytic sites with lateral interactions and with participation of the support are expected to be hardly possible. In such situations, the kinetic Monte Carlo simulation may be efficient [2][3][4]36].…”

“…Here, we analyze the case when CO adsorbs on one sites (S1) and O 2 and O adsorb on the other sites (S 2 ), and the reaction occurs via steps (1)- (4). For this model, the conventional MF equations are as follows…”

Kinetic modelling of heterogeneous catalytic systems

“…Limitations in the accuracy of DFT can also lead to qualitative deviations from experimental data, exemplified for instance by the so-called 'CO puzzle': the incorrect prediction (by semi-local DFT functionals) that CO would preferentially bind to hollow, rather than top sites [231][232][233] (for a recently developed correction method see [234]). In turn, such qualitative and quantitative deviations may result in unreliable predictions: as demonstrated by recent studies of the CO oxidation on RuO 2 (1 1 0), slightly different parameter sets for a KMC model thereof may lead to distinctly different catalytic behaviours [235,236]. These observations underscore not only the needs for more accurate DFT methods and the validation of the KMC results by detailed comparisons with experimental data [236], but also the need for methods that can quantify the effect of the uncertainty of certain parameters on the simulation results.…”

“…In turn, such qualitative and quantitative deviations may result in unreliable predictions: as demonstrated by recent studies of the CO oxidation on RuO 2 (1 1 0), slightly different parameter sets for a KMC model thereof may lead to distinctly different catalytic behaviours [235,236]. These observations underscore not only the needs for more accurate DFT methods and the validation of the KMC results by detailed comparisons with experimental data [236], but also the need for methods that can quantify the effect of the uncertainty of certain parameters on the simulation results. Besides the need for uncertainty quantification, in complex systems it is also essential to identify the main parameters influencing…”