The sputtering of Mercury surface analogues in models and experiments

Surface sputtering, the process of energetic ions (e.g. from the solar wind) ejecting  particles from the surface, is one of the major processes to supply energetic particles to a thin, collisionless atmosphere on otherwise atmosphere-free celestial bodies. The energetic ejecta reach high altitudes or may escape the celestial body completely so they can be observed remotely from the ground or in-situ with a spacecraft. The accuracy of a model to reproduce the thin atmosphere surrounding exposed rocky bodies such as the Moon and Mercury therefore requires accurate sputter efficiencies.

Determining the sputter yields of the various species from a realistic mineral surface is still a work in progress [1]. Modeling of sputtering with commonly used Binary Collision Approximation (BCA) models such as TRIM [2] has been shown to overestimate the sputter yields compared to experimental data [3, 4]. The number of sputter experiments performed on rock forming minerals is growing steadily, however. We apply the latest results to obtain yields for a range of minerals from the state-of-the-art code SDTrimSP [5], which is an improved model based on the static TRIM and the dynamic TRIDYN [6]. 

We present a novel implementation of oxides as components to appropriatly represent mineral densities in SDTrimSP by distinguishing between bound and free O atoms in the simulation. We compare the oxide implementation to recent sputter yields obtained from mineral pellets and mineral-derived thin-film irradiation experiments [2,3,4]. One of our preliminary results is shown in Figure 1 where we compare the SDTrimSP sputter yields of the commonly used atomic model and the newly implemented oxide model with experimental data from [4]. For reference we also show the atomic model equivalent results from the widely used SRIM code. This work includes a parameter study including density, binding energy, and binding model and how they affect the resulting mineral sputter yields as well as the ejectas energy and angular distribution. We further elaborate on necessary experiments to better constrain the parameter spaces. 

Figure 1: Sputter yields of solar wind energy hydrogen ions on wollastonite (CaSiO3). SDTrimSP results for the default atomic model and the newly implemented oxide model recreate experimental data from [4] reasonably well compared to the commonly used SRIM.

[1] Jäggi, N., et al. (2021). Icarus, 365, 114492. 

[2] Ziegler, J.F., et al. (2010). Nucl. Instrum. Methods Phys. Res. B, 268, 1818–1823. 

[3] Biber H., et al. (2020). Nucl. Instrum. Methods Phys. Res. B, 480, 10. 

[4] Szabo, P.S., et al. (2018). Icarus, 314, 98–105.

[5] Mutzke, A., et al. (2019). SDTrimSP Version 6.00. Max-Planck-Institut für Plasmaphysik.

[6] Möller, W., & Eckstein, W. (1984). Nucl. Instrum. Meth. B., 2(1–3), 814–818