Ionic liquid binary mixtures: Machine learning‐assisted modeling, solvent tailoring, process design, and optimization

Abstract

This work conducts a comprehensive modeling study on the viscosity, density, heat capacity, and surface tension of ionic liquid (IL)‐IL binary mixtures by combining the group contribution (GC) method with three machine learning algorithms: artificial neural network, XGBoost, and LightGBM. A large number of experimental data from reliable open sources is exhaustively collected to train, validate, and test the proposed ML‐based GC models. Furthermore, the Shapley Additive Explanations technique is employed to quantify the influential factors behind all the studied properties. Finally, these ML‐based GC models are sequentially integrated into computer‐aided mixed solvent design, process design, and optimization through an industrial case study of recovering hydrogen from raw coke oven gas. Optimization results demonstrate their high computational efficiency and integrability in solvent and process design, while also highlighting the significant potential of IL‐IL binary mixtures in practical applications.

Interpretable Feedforward Neural Network and XGBoost-Based Algorithms to Predict CO2 Solubility in Ionic Liquids

“…Within the AI domain, machine learning (ML) methods have gained traction, especially in property estimation for various compounds. ML techniques typically require numerical vector representations or nominal values for learning examples, with quantitative structure–property relationship (QSPR) models providing a robust framework for such representations. − According to our literature survey, some examples of commonly used ML techniques include support vector machine regression (SVM), multilayer perceptron (MLP), decision tree regression (DT), random forest regression (RF), gradient boosting regression (GB), extreme gradient boosting regression (XGBoost), neural network (NN), and etc.…”

Modeling Study on the Density and Viscosity of Ionic Liquid–Organic Solvent–Water Ternary Mixtures

“…25 Additionally, hybrid modeling methods that combine GC methods and ANN algorithms have been introduced to model the phase equilibrium behavior of aqueous two-phase systems (ATPSs). 26,27 Recent works have also demonstrated the seamless integration of ANN-based GC models with computer-aided design methods for the optimal design of solvent 13 and aqueous biphasic systems. 28−30 Inspired by these successful applications, this study aims to combine the GC method with three popular ML algorithms, namely, ANN, XGBoost, and LightGBM, to build models for predicting the density and viscosity of IL−os−water mixtures.…”

“…To the best of our knowledge, most of the published works mainly focus on the property prediction of pure or binary mixtures of ILs, − and modeling studies on the ternary mixtures containing ILs, especially IL–os–water mixtures, are still scarce. As reported, IL–os–water mixed solvents often exhibit superior performance compared to pure ILs in many cases.…”

“…In our previous work, we studied the density, viscosity, and other properties of the IL−IL binary hybrid system and concluded that ANN and GC have a good combination. 13 However, Shu et al used a similar method to study the surface tension properties of the IL−os binary mixed system and concluded that the combination of XGB and GC had a better prediction effect than that of ANN and GC. 15 It can be seen that the combination of ANN and GC will not necessarily have the best predictive effect on any ionic liquid mixing system.…”

“…Recently, machine learning (ML) algorithms such as artificial neural networks (ANNs) and support vector machine (SVM) have been employed to develop nonlinear GC models for predicting CO 2 solubility in ionic liquids, toxicity of ionic liquids, and physical properties of IL–H 2 O mixtures such as viscosity and surface tension . Additionally, hybrid modeling methods that combine GC methods and ANN algorithms have been introduced to model the phase equilibrium behavior of aqueous two-phase systems (ATPSs). , Recent works have also demonstrated the seamless integration of ANN-based GC models with computer-aided design methods for the optimal design of solvent and aqueous biphasic systems. − Inspired by these successful applications, this study aims to combine the GC method with three popular ML algorithms, namely, ANN, XGBoost, and LightGBM, to build models for predicting the density and viscosity of IL–os–water mixtures. In our previous work, we studied the density, viscosity, and other properties of the IL–IL binary hybrid system and concluded that ANN and GC have a good combination .…”