Study of the adsorption of light species on interstellar ices using molecular dynamics and machine-learned potentials

Modeling the chemistry happening on top of interstellar dust grains requires information on several properties (sticking coefficients, binding energies, etc.) of adsorbates on surfaces. With the conditions of dense clouds, solid bodies are found as amorphous ices, where the quantities mentioned above are not univocally defined due to the distribution of binding sites on the surface. Therefore, these quantities' theoretical determination requires sampling of the expected outcomes on the surface's different sites. Here, we present our recent results on the study of these adsorption processes by combining sampling using ab-initio molecular dynamics simulations with the help of cheap electronic structure solvers [1]. We will focus our attention on using machine-learned potentials [2, 3] to affordably tackle these problems.

[1] Molpeceres, G., & Kästner, J. (2020). Phys. Chem. Chem. Phys., 22(14), 7552–7563.
[2] Zaverkin, V., & Kästner, J. (2020). J. Chem. Theory Comp., 16(8), 5410–5421.
[3] Molpeceres, G., Zaverkin, V., & Kästner, J. (2020). Mon. Not. Royal Astron. Soc., 499(1), 1373–1384.