Computer-Aided Design of New Physical Solvents for Hydrogen Sulfide Absorption

Treatment of hydrogen sulfide (H2S) is important in many industrial processes including oil refineries, natural and biogas processing, coal gasification. The most mature technology for the selective capture H2S based on its absorption by chemical or physical solvents. However, only several compounds are currently used as (co)solvents industry, search new ones an task. experimental screening requires much time resources, while solubility modeling might enable one to reduce number solvents evaluation. In this study, a workflow silico discovery was suggested experimentally validated. A data set composed 99 collected predictive quantitative structure–property relationships were built using random forest algorithm two types molecular descriptors: ISIDA fragments quantum-chemical descriptors. Virtual industrially produced chemicals their structural analogues enabled identification with predicted high values. They can be starting points further exploration space. value found virtual screening, 1,3-dimethyl-2-imidazolidinone, confirmed experimentally.