Login / Signup

Like dissolves like: A first-principles theory for predicting liquid miscibility and mixture dielectric constant.

Bilin ZhuangGabriele RamanauskaiteZhao Yuan KoaZhen-Gang Wang
Published in: Science advances (2021)
Liquid mixtures are ubiquitous. Miscibility and dielectric constant are fundamental properties that govern the applications of liquid mixtures. However, despite their importance, miscibility is usually predicted qualitatively based on the vaguely defined polarity of the liquids, and the dielectric constant of the mixture is modeled by introducing mixing rules. Here, we develop a first-principles theory for polar liquid mixtures using a statistical field approach, without resorting to mixing rules. With this theory, we obtain simple expressions for the mixture's dielectric constant and free energy of mixing. The dielectric constant predicted by this theory agrees well with measured data for simple binary mixtures. On the basis of the derived free energy of mixing, we can construct a miscibility map in the parameter space of the dielectric constant and molar volume for each liquid. The predicted miscibility shows remarkable agreement with known data, thus providing a quantitative basis for the empirical "like-dissolves-like" rule.
Keyphrases
  • ionic liquid
  • electronic health record
  • big data
  • high resolution
  • machine learning
  • artificial intelligence
  • deep learning
  • data analysis