| Due to their very attractive physical properties, such as wide liquid-phase range, excellent solubility, negligible vapor pressure, and designability, ionic liquids (ILs) are widely considered as a novel solvent in separation processes. However, as a designer solvent, there exists a large library of ILs, which causes challenges to quickly find proper ILs for a specific separation problem. In order to solve this problem, a computer-aided ionic liquid design method based on multi-scale simulations is presented.For the solution of the computer-aided IL solvents design (CAILD) problem with COSMO-SAC model, the group contribution method was adopted in this work. And thus, ILs were decomposed into three parts:cation skeleton, substitutes and anion. By regressing the surface charge density (?-profile) and cavity volume of 828 cations and 61 anions, the COSMO information of groups were acquired, and a method called GC-COSMO was proposed to predict the ?-profile and VCOSMO of the ILs.To ensure the designed IL solvents all meet the practical application, the structural and physical property constraints were added to the design procedure, in which the physical properties, such as melting point, viscosity and surface tension, were calculated by semi-empirical models. The design problem was then formulated as a mixed integer nonlinear programming (MINLP) problem by incorporating all the prediction models and constraints using MATLAB, where the objective function was set to the separation performance of the IL solvents. To solve the MINLP problem, a deterministic algorithm called branch and bound was used.Two separation tasks of benzene-cyclohexane extraction and post-combustion carbon dioxide capture were tested to verify the CAILD method. Since the design results were highly consistent with literature, the proposed design method was thought to be efficient and correct for designing the proper IL solvents for separation process. |
PDF Full Text Request