John C. Tully
Abstract:The rates and pathways of chemical reactions at metal surfaces can be strongly influenced by energy dissipation due to the nonadiabatic excitation of metallic conduction electrons. The introduction of frictional forces to account for this dissipation has been quite successful in situations for which the nonadiabatic coupling is weak. However, in cases where nonadiabatic coupling is strong, such as when electron transfer occurs, the friction model is likely to break down. Ryabinkin and Izmaylov have proposed 2-state and 3-state alternatives to the friction model for introducing electronic dissipation in molecular dynamics simulations. Here, we examine their 2-state model using some simple examples of atom–surface scattering. We find that, with the addition of decoherence, the 2-state model can produce quite promising results.
What problem does this paper attempt to address?
The paper primarily discusses the issue of energy dissipation during chemical reactions on metal surfaces, with a particular focus on the influence of non-adiabatic effects, such as the excitation of metal conduction electrons, on reaction rates and pathways. Traditionally, to describe this energy dissipation, the concept of frictional forces has been introduced in research, and this approach has been successful in cases of weak non-adiabatic coupling. However, the friction model may fail in situations of strong non-adiabatic coupling (for example, during electron transfer).
To address this limitation, Ryabinkin and Izmaylov proposed two alternative schemes—the two-state model and the three-state model—to better simulate the electron dissipation during the interaction of molecules with metal surfaces. This paper focuses on the two-state model and validates its effectiveness through some simple examples of atomic-surface scattering.
The authors found that, with the inclusion of decoherence effects, the two-state model could produce promising results. By studying three different scenarios (including one that fits the conditions where the friction model is applicable and two that do not), the results show that, with the proper choice of decoherence parameters, the two-state model can reproduce the results obtained by the Ehrenfest method, even in cases of strong coupling or the presence of resonant states.
Therefore, the core issue that the paper attempts to address is how to effectively simulate the dynamic behavior of molecules on metal surfaces in situations of strong non-adiabatic coupling, especially where electron transfer may occur. The two-state model offers a new, simple method to handle such problems, demonstrating better accuracy under a broader range of conditions compared to the traditional friction model.