(383ay) Development of a Generalized Gradient Boosting Regression Framework for Predicting IL System Properties.

Ionic liquids (ILs) are a novel class of salts that exist as liquids at < 100℃ and exhibit several favorable properties, including high conductivity, low volatility, and high thermal and electrochemical stability. In recent years, IL systems - such as pure ILs, binary ILs, and IL-solvent mixtures - have demonstrated great potential for a range of electrochemical applications, including electrolytes for non-lithium batteries, supercapacitors, dye-sensitized solar cells, and thermo-electrochemical cells^1–3. However, the vast design space of ILs makes experimental characterization of all possible IL systems infeasible, hindering the development of new IL systems. For instance, there are over one million pure ILs and more than possible binary IL mixtures, making exhaustive experimentation impractical^4,5. This highlights the need for cost-effective, high-throughput computational methods to predict the properties of IL systems in order to harness their versatile design space.

In this work, we employ computational and data analysis techniques to predict the thermophysical properties of single, binary solvent-IL systems. We present a computational framework built on a Gradient Boosting Regression (GBR) algorithm, capable of making generalized density and viscosity predictions for a wide range of IL systems, using features derived from COSMO-RS-generated sigma profiles. We demonstrate the framework's capability to predict both density and viscosity for various IL systems and to capture key property relationships, such as temperature and composition dependencies, observed in experimental data and present important insights for the computer aided design of IL systems.

References

(1) Lall-Ramnarine, S.; Suarez, S.; Zmich, N.; Ewko, D.; Ramati, S.; Cuffari, D.; Sahin, M.; Adam, Y.; Rosario, E.; Paterno, D.; Wishart, J. Binary Ionic Liquid Mixtures for Supercapacitor Applications. ECS Trans. 2014, 64 (4), 57–69. https://doi.org/10.1149/06404.0057ecst.

(2) Kaur, G.; Kumar, H.; Singla, M. Diverse Applications of Ionic Liquids: A Comprehensive Review. Journal of Molecular Liquids. Elsevier B.V. April 1, 2022. https://doi.org/10.1016/j.molliq.2022.118556.

(3) Macfarlane, D. R.; Tachikawa, N.; Forsyth, M.; Pringle, J. M.; Howlett, P. C.; Elliott, G. D.; Davis, J. H.; Watanabe, M.; Simon, P.; Angell, C. A. Energy Applications of Ionic Liquids. Energy and Environmental Science. Royal Society of Chemistry 2014, pp 232–250. https://doi.org/10.1039/c3ee42099j.

(4) Chatel, G.; Pereira, J. F. B. B.; Debbeti, V.; Wang, H.; Rogers, R. D. Mixing Ionic Liquids-"simple Mixtures" or “Double Salts”? Green Chemistry. Royal Society of Chemistry 2014, pp 2051–2083. https://doi.org/10.1039/c3gc41389f.

(5) Niedermeyer, H.; Hallett, J. P.; Villar-Garcia, I. J.; Hunt, P. A.; Welton, T. Mixtures of Ionic Liquids. Chem. Soc. Rev. 2012, 41 (23), 7780–7802. https://doi.org/10.1039/c2cs35177c.

Research Interests

AI, ML, and Data Science for Molecules and Materials