A Machine Learning approach for enhancing permittivity mixing rules of binary liquids with a Gaussian modification and a new interaction factor estimation

M. Monteagudo Honrubia, F.J. Herraiz-Martínez, J. Matanza

Journal of Molecular Liquids Vol. 399, pp. 124290-1 - 124290-14

Original summary:

The microstructure and solvation mechanics of binary liquids are key for predicting mixture permittivity. However, since traditional mixing rules do not consider this complexity, they must be modified to address the mixture characteristics through an interaction factor (Κ_int). This paper evaluates this parameter for several mixing rules, applying Support Vector Regressor models trained with glycerin-water reflective signals acquired with a Dielectric Resonator sensor. The regression error of these models indicates both the optimal interaction factor and the mixing rule that fits the most with experimental permittivity values. Kraszewski and Hashin-Shtrikman mixing rules achieved the best performance with an RMSE of around 1. In addition, this paper suggests that the interaction factor can be estimated through the molar volume and the dielectric contrast between liquids (K_int=2.67) without acquiring experimental data. Moreover, after analyzing the physical limitations of a linear modification formula, this paper proposes an alternative based on a Gaussian function that avoids unrealistic volume fractions. Both contributions enhance mixing rule accuracy and improve the flexibility to model mixture dielectric behavior.

English summary:

The microstructure and solvation mechanics of binary liquids are key for predicting mixture permittivity. However, since traditional mixing rules do not consider this complexity, they must be modified to address the mixture characteristics through an interaction factor (Κint). This paper evaluates this parameter for several mixing rules, applying Support Vector Regressor models trained with glycerin-water reflective signals acquired with a Dielectric Resonator sensor. The regression error of these models indicates both the optimal interaction factor and the mixing rule that fits the most with experimental permittivity values. Kraszewski and Hashin-Shtrikman mixing rules achieved the best performance with an RMSE of around 1. In addition, this paper suggests that the interaction factor can be estimated through the molar volume and the dielectric contrast between liquids (K_int=2.67) without acquiring experimental data. Moreover, after analyzing the physical limitations of a linear modification formula, this paper proposes an alternative based on a Gaussian function that avoids unrealistic volume fractions. Both contributions enhance mixing rule accuracy and improve the flexibility to model mixture dielectric behavior.

Spanish layman's summary:

Las mixing rules deben modificarse para considerar las características de la mezcla a través de un factor de interacción, que puede ser evaluado para varias mixing rules mediante el error de regresión de los modelos ML. además, este parámetro puede estimarse a través del volumen molar y el contraste dieléctrico entre líquidos. Después de analizar las limitaciones de la fórmula de modificación lineal, este artículo propone una alternativa basada en una función gaussiana que evita fracciones de volumen poco realistas. Ambas contribuciones mejoran la precisión de las mixing rules y mejoran la flexibilidad para modelar el comportamiento dieléctrico de la mezcla.

English layman's summary:

Traditional mixing rules must be modified to address the mixture characteristics through an interaction factor, which can be evaluated for several mixing rules through the regression error of ML models. In addition, this paper suggests that this parameter can be estimated through the molar volume and the dielectric contrast between liquids. After analyzing the physical limitations of a linear modification formula, this paper proposes an alternative based on a Gaussian function that avoids unrealistic volume fractions. Both contributions enhance mixing rule accuracy and improve the flexibility to model mixture dielectric behavior.

Keywords: Binary mixture; Dielectric characterization; Machine Learning; Mixing rules; Permittivity

JCR Impact Factor and WoS quartile: 6,000 - Q1 (2022)

DOI reference: https://doi.org/10.1016/j.molliq.2024.124290

Published on paper: April 2024.

Published on-line: February 2024.

Citation:
M. Monteagudo Honrubia, F.J. Herraiz-Martínez, J. Matanza, A Machine Learning approach for enhancing permittivity mixing rules of binary liquids with a Gaussian modification and a new interaction factor estimation. Journal of Molecular Liquids. Vol. 399, pp. 124290-1 - 124290-14, April 2024. [Online: February 2024]