ABSTRACT Forsterite is a primary constituent of interstellar dust and planetary systems. It is believed to originate from the outflows of oxygen-rich stars and undergo further processing within the interstellar medium through the action of cosmic rays and shocks. Under these harsh conditions, point defects may form, such as MgO Schottky vacancies. These vacancies can then undergo atom reconstruction as part of a chemical process to maintain the system’s crystalline structure. Polycyclic aromatic hydrocarbons (PAHs) are ubiquitously observed interstellar molecules and are thought to form through gas-phase reactions akin to sooting flames. However, their role and impact on dust stability remain unknown. In this study, we employ an atomistic artificial-intelligence-based method, surrogate machine learning trained directly by density functional theory. Specifically, we utilize gofee (global optimization with first-principles energy expressions) to predict possible reconstructions of MgO vacancies on a crystalline forsterite (010) surface as an important component of interstellar dust and planetary systems. We identify nine possible reconstructions involving the formation of unbound Si and O atoms. We investigate their energy stability and find that the reconstruction of Si–O atoms stabilizes the vacancy by about 0.54 eV. Additionally, if PAHs bind with the unbound O and Si atoms of the vacancy, the vacancy is stabilized by approximately 1.76 eV. We demonstrate that PAHs, along with the reconstruction of unbound atoms on the dust surface, affect the stability of the dust, which might open up avenues for diverse chemistry.
Read full abstract