Bridging Machine Learning and Redlich–Kister Theory for Solid–Liquid Equilibria Prediction of Binary Eutectic Solvent Systems

Abstract

Eutectic solvents (ESs) have gained significant interest in various chemical processes due to a broad spectrum of attractive properties, whereas their rational design is currently still in its infancy. To bridge this gap, Redlich−Kister (RK) theory and machine learning are linked for the solid−liquid equilibria (SLE) prediction of ES systems, which is thermodynamically the cornerstone for ES design. RK theory with two or three parameters is first evaluated by fitting experimental SLE of an extensive ES database, demonstrating that the two-parameter-based one is sufficiently reliable for eutectic behavior correlation. Three machine learning methods, namely, Random Forest, multiple linear regression, and ElasticNet, are developed for relating the parameters of RK theory to the RDKit descriptors of ES components. The SLE predictions from RK theory parametrized by the developed machine learning models are carefully evaluated and further externally examined on several recently reported ES systems.

DESignSolvents: an open platform for the search and prediction of the physicochemical properties of deep eutectic solvents

“…31 Regarding the melting temperature, existing models are also trained on limited datasets primarily associated with specific types of DESs. 21,41,42 In this work, we propose more general machine learning models designed to predict the aforementioned physicochemical characteristics for various types of binary and ternary DESs. Fig.…”

Computer-Aided Molecular Design of Ionic Liquids as Advanced Process Media: A Review from Fundamentals to Applications

“…93 The finally obtained model comprising 14 molecular descriptors shows a high predictive performance with a MAE of 13.89 K for T e and 0.0973 for x e . Very recently, Wang et al 533 further improved the performance of this model by employing the Redlich−Kister (RK) theory and ML methods (RF and ElasticNet). Rigorous internal and external cross-validations were carried out to develop the model (Figure 33).…”

“…Through the comparative screening, it was indicated that ESs could be considered as a superior option to ILs for the physical absorption of CO 2 . It is worth mentioning again that the RS/RK-ML models developed by Chen et al 532 and Wang et al 533 also allow the reverse design of novel ESs by estimating the SLE behaviors of different component combinations.…”

“…Figure 33. Framework for developing ensemble and linear models adopted by Wang et al533 Copyright 2023 American Chemical Society.…”

A Systematic Review on Intensifications of Artificial Intelligence Assisted Green Solvent Development

“…In addition, more and more supporting databases and software are being developed and continuously improved. Up to now, QSPR has evolved from simple regression analysis to a multiple statistical ML technique, which can analyze a large scale of chemical structures. − Being able to simulate the physical, chemical, and biological properties, − ML-based QSPR models have been widely used to guide the development of green solvents (Table ).…”