Heterogeneous catalysis is a key technology in modern society. To develop new catalysts, it is essential to understand chemical reactions on catalyst surfaces. Researchers have studied these interactions extensively using both experiments and theory. Among theoretical methods, density functional theory (DFT) is a powerful tool for analyzing how molecules adsorb on metal surfaces, including their structures and energies.

In the field of chemical reaction on catalysts, the role of the dispersion forces in adsorption on metal surfaces is important. The dispersion forces arise from fluctuating dipole moments of atoms and molecules, yielding the long-range attractive force of the van der Waals (vdW) interaction. The fluctuations are taken into account in the new type of the correlation functionals called van der Waals functional. Recently, the augmentation of the generalized gradient approximation (GGA) with the vdW functional (vdW correction) was found to be essential not only for physisorbed systems but also for chemisorbed molecules. In addition, ab initio molecular dynamics (AIMD) simulation enables tracking catalytic reactions on the femtosecond scale using quantum mechanical mechanics. While this method can reveal detailed reaction mechanisms and dynamics, its accuracy depends heavily on the choice of functionals. Therefore, selecting suitable functionals is increasingly important for modeling specific catalytic processes accurately.

Here, we investigated the CO oxidation and CO₂ desorption processes on Pt(111) through density functional theory (DFT) calculations employing three different exchange-correlation functionals, namely PBE, vdW-DF, and optB86b-vdW. AIMD simulation was performed to elucidate the desorption dynamics of CO₂ on Pt(111). The oxidation reaction of CO on Pt surfaces is a prototype catalytic reaction and has long been investigated both experimentally and theoretically. When the CO+O on the Pt(111) surface is heated to 200–350 K, these adsorbates are associated to desorb as ${\beta }_1$, ${\beta }_2$, and ${\beta }_3$ CO₂ states [1]. For the desorbed CO₂ molecules, the dynamic properties including kinetic energy, angular distribution, and vibrational states have been investigated [2-5]; a hyperthermal velocity distribution and a narrow desorption angular distribution has been reported and they are assigned to the formation and desorption of CO₂ from the Pt(111) terrace [5]. However, the previous DFT calculation could not correctly reproduce the experimental results including the kinetic energy of CO₂ (0.38 eV) and ${\cos }^8\theta$ desorption angular distribution [5], probably because their functional did not include the vdW corrections. Since the vdW correction is necessary for correctly calculating the adsorption energy of CO₂ on the Pt(111) surface, it would be crucial for a deeper understanding of the CO oxidation kinetics and dynamics on the Pt(111) surface.

Our calculated potential energy surfaces [7] indicate that the kinetically favorable pathway for the CO oxidation is the hexagonal close packed (HCP) pathway, where the oxidation is initiated by the movement of an O atom at the HCP site toward an adsorbed CO at the on-top site (Fig. 1).

The present AIMD simulation [7] has also revealed that “vdw-DF” functional, provides the most quantitative explanation of the experimental results regarding the distribution of the kinetic energy, the sharp desorption angle, and the OCO bending vibration of the desorbed CO₂. The good agreement is the result of an accurate description of the reactant (CO + O) and product (CO₂) in both the chemisorbed and physisorbed states on Pt(111). The present AIMD simulation also clearly shows that the CO₂ is desorbed in vibrationally excited states with the energy transfer between the bending and symmetric stretching modes (Fermi resonance).

References

• [1] J. Yoshinobu and M. Kawai, J. Chem. Phys. 103, 3220 (1995).
• [2] T. Matsushima, Surf. Sci. 127, 403 (1983).
• [3] K. H. Allers et al., J. Chem. Phys. 100, 3985 (1994).
• [4] K. Nakao et al., J. Phys. Chem. B 109, 24002 (2005).
• [5] J. Neugebohren et al.., Nature, 558, 280 (2018).
• [6] L. Zhou, Angew. Chem. Int. 131, 6990 (2019).
• [7] H. Li et al., J. Phys. Chem. C 128, 15393 (2024).

Authors

• H. Li, Y. Kataoka, S. Tanaka, J. Haruyama, O. Sugino, and J. Yoshinobu