查看文章

Deep eutectic solvents (DESs) were discovered as a new class of green solvents and attracted high interest of researchers in various fields of science. Multiple hydrogen bonding between DES components leads to formation of eutectic mixture with significantly lower melting point in comparation with initial compounds [1]. Currently, new DESs are developing by experimental approach that consumes time and resources [1, 2]. To the best of our knowledge, there is no general theoretical method presented in the literature of DES components selection or eutectic point prediction algorithm successfully applied to DESs. Hence, the aim of our research is to identify the limitations and possible opportunities of the eutectic point prediction methods in context of DESs based on machine learning approach.

The most relevant method for eutectic point prediction is COSMO-RS model that allows solid-liquid equilibrium modelling …