Predicting the ET(30) parameter of organic solvents via machine learning

Polarity of organic solvents is an important parameter which needs to be considered during a reaction design as it can drastically impact the rate and dynamics of a chemical reaction. Till now ET(30) scale is the only comprehensive scale which can accurately quantify various solute–solvent and solvent–solvent interactions, the experimental determination of which is an expensive and resource-intensive approach. Therefore, we have resorted to machine learning techniques for predicting the empirical polarity of organic solvents which would provide ET(30) values for new solvents in a fast and efficient manner without having to rely on experimental and computational setup.