Seminar:

Clark Hall Room 214

Gas-surface reactions play an important role in many heterogeneous catalysis processes such as methane steam reformation and water-gas shift reactions. An in-depth understanding of these dynamical processes is of great importance and requires a completely dynamical model. However, most of previous theoretical studies have neglected the degrees of freedom of the surface atoms, i.e. within the static surface approximation. Recently, we have developed high-dimensional potential energy surfaces (PESs) including surface atoms for describing the molecule-surface energy exchange, taking advantage of the Behler-Parrinello (BP) type of atomistic neural network (AtNN) method.¹ Taking CO₂/Ni(100) systems as an example, we show that molecular dynamics simulations on NN PES reproduce well the much more expensive on-the-fly ab-initio molecular dynamics (AIMD) results with much better statistics.^{2, 3} More importantly, we can obtain a dynamically converged high-dimensional PES with as few as fifty AIMD trajectories, which enable us to predict more demanding state-to-state scattering properties of polyatomic molecules on metal surfaces.⁴ I will also show some preliminary results for the NO scattering from Au(111) system, for which the importance of an adiabatic PES is emphasized even though the dynamics is apparently non-adiabatic.⁵ If there is sufficient time, I will briefly talk about our new type of AtNN model, physically inspired from the well-known embedded atom method. Our new implementation is much faster than the BP type of AtNN framework with the same level of accuracy.⁶

¹ J. Behler, and M. Parrinello, Phys. Rev. Lett. 98, 146401 (2007).

² Q. Liu, X. Zhou, L. Zhou, Y. Zhang, X. Luo, H. Guo, and B. Jiang, J. Phys. Chem. C 122, 1761 (2018).

³ B. Kolb, X. Luo, X. Zhou, B. Jiang, and H. Guo, J. Phys. Chem. Lett. 8, 666 (2017).

⁴ Y. Zhang, X. Zhou, and B. Jiang, J. Phys. Chem. Lett. 10, 1185 (2019).

⁵ R. Yin, Y. Zhang, and B. Jiang (submitted)

⁶ Y. Zhang and B. Jiang (in prepration)