Data-Driven Modelling of Gas Solubility in Ionic Liquids Using Principal Properties as Orthogonal Descriptors

摘要

Principal properties for gases and ionic liquids (ILs) were used as descriptors to model gas solubility in ILs by means of datadriven approaches such as partial least squares (PLS) and orthogonal PLS (OPLS). Satisfactory PLS predictions, spanning above 3 log units, can be easily achieved using a nine parameters equation, while disjoint OPLS models for different gases, namely hydrocarbons, neutral and acidic ones, allow more accurate predictions.