Ab initio calculation of binding energies of sulphur containing astrochemical relevant species with crystalline interstellar ice models

In the interstellar medium (ISM), i.e. the matter filling the space between the stars, it is possible to define several environments, depending on density, temperature, and chemical composition. Among all of them, the molecular clouds often referred to as the cradle of stellar birth, are the objects of this thesis. In the molecular clouds, radio to far-infrared observations revealed several molecules in the gas phase, while near-infrared spectroscopy detected the presence of sub-micron sized dust grains covered by multiple layers of H2O-dominated ices, often called “dirty ice” due to the presence of other volatile molecules. The strength of the interaction between species in the gas phase and the surface of the ice is measured by the binding energy (BE), a crucial parameter in the astrochemical numerical models which describe the evolution of the ISM chemistry. In this work, we focused on 17 sulfur-containing species to investigate the depletion of sulfur, a long-standing issue in the field of astrochemistry. Indeed, each element of the periodic table is expected to be present in the universe in a certain amount, called cosmic abundance, which has been estimated from the composition of the Solar System. While in diffuse regions of the ISM sulfur is mostly present in its ionized atomic form, in molecular clouds and star-forming regions its cosmic abundance is dramatically reduced in the gas-phase and it is puzzling what can be the reservoir of the element. One possibility is that S-containing species remain bound to the ice mantle and therefore the purpose of this thesis is to quantify this possibility through the calculation of the BE for the considered species. To that purpose, each S-containing species was adsorbed on the (010) surface of a proton-ordered crystalline ice model. Density functional theory (DFT), based on the B3LYP-D3(BJ), M06-2X, and MPWB1K-D3(BJ) functionals, was used for the prediction of the structures and energetics. In some cases, more than one adsorption complex was found for each molecule. The DFT BEs were also refined by adopting an ONIOM-like procedure to estimate the correction at CCSD(T) level with basis set extrapolation. We found that the plain DFT BEs are very close to those corrected at CCSD(T) level. As the time-limiting step in this study was the geometry optimization of the species adsorbed at the ice surface we developed a computationally cheap but accurate recipe to arrive at the BE values of the same quality of those computed at full DFT level. We adopted the computationally fast HF-3c method to optimize the geometries, while the final BE are obtained by a single point energy calculations at the DFT//HF-3c level, i.e. by adopting the HF-3c optimized structures. Comparison between DFT//HF-3c and full DFT BE values showed a very good linear correlation between the two sets of data, validating the proposed procedure. Our computed data were compared with the available literature experimental data, showing significant differences in the majority of the cases. The behavior of S-containing species was compared with their O-containing analogous, when possible, and with a large set of 21 astrochemical relevant molecules previously studied in a precedent MSc thesis by Stefano Ferrero.