Many of us have experience in modelling heterogeneous catalysts using mean-field-type approaches, e.g. within the context of Langmuir-Hinshelwood kinetics. However, an increasing amount of research is showing that such approaches are inadequate, at least for certain systems. The approximations and simplifications employed therein can indeed lead to quantitative or even qualitative errors in the prediction of catalytic performance metrics. This tutorial is intended for readers with background on chemical kinetics but who have just started exploring kinetic Monte Carlo (KMC), and discusses what this method can do and why it is superior to mean-field-type models.
If you are wondering "What can Kinetic Monte Carlo (KMC) do for me?" this article is for you. It is intended as a discussion more than a tutorial in the strict sense, but with a (hopefully) clear learning aspect. Thus, we will highlight the benefits of KMC compared to simpler approaches, such as mean-field (micro-kinetic or phenomenological) models, taking as an example a prototypical system: the CO oxidation reaction. Of course, the main ideas behind this discussion hold true for more complicated reactions as well.
So the net reaction we will discuss is:
CO + ½O2 → CO2
When this happens on the surface of a catalyst (e.g. the (111) facet of a Pd nano-crystal) the reaction could proceed via the following three elementary steps:
CO + * ↔ CO* (reversible adsorption of CO on an empty catalytic site)
O2 + 2* ↔ 2O* (reversible dissociative adsorption of O2 on two neighbouring empty catalytic sites)
CO* + O* → CO2 + 2* (reversible adsorption of CO on an empty catalytic site)
In the notation we just used and will adopt throughout, * denotes an empty site, whereas A* an adsorbate (bound species A).