For discovering uncharted chemical space of ionic liquids (ILs) for CO₂ dissolution, a reliable generative framework combining re-balanced variational autoencoder (VAE), artificial neural network (ANN), and particle swarm optimization (PSO) is developed based on a comprehensive experimental solubility database from literature. The re-balanced VAE transforms the chemical space of ILs into continuous latent space, which is demonstrated by t-distributed stochastic neighbor embedding (t-SNE) visualization and sampled ions of the latent space. ANN is connected with the re-balanced VAE to predict the CO₂ solubility and the resultant VAE-ANN model achieves a low mean absolute error (MAE) of 0.022 on the test set. Lastly, the PSO algorithm is employed to search the latent space for optimal IL structures with the highest predicted solubility. A total of 5120 ILs are generated and optimized through 10 parallel runs of PSO. Their CO₂ solubilities are predicted and compared to those of the 3735 ILs combined with the already-known cations and anions in the CO₂ solubility database under 298.15 K and 100 kPa. The results demonstrate a notably larger distribution of higher CO₂ solubility in optimized ILs after PSO, which effectively points out the significance and directions for exploring the wide IL chemical space.

When considering the usage of ionic liquids (ILs) for reactions and separations involving non-polar or weak-polar hydrocarbons, the knowledge of the mutual solubility behaviors of ILs and hydrocarbons is of the utmost importance. In this work, taking four typical C6-hydrocarbons namely benzene, cyclohexene, cyclohexane, and hexane as representatives, the mutual solubility of ILs and non-polar or weak-polar hydrocarbons are systematically studied based on the COSMO-RS model. The reliability of COSMO-RS for these systems is first evaluated by comparing experimental and predicted hydrocarbon-in-IL activity coefficient at infinite dilution and binary/ternary liquid-liquid equilibria of related systems. Then, the mutual solubility of the four hydrocarbons and 13,650 ILs (composed by 210 cations and 65 anions) are predicted. The effect of different IL structural characteristics including alkyl chain length, cation family/symmetry/functional group, and anion on the IL-hydrocarbon mutual solubility behaviors are further analyzed by the analyses of interaction energy and screen charge distribution. The mutual solubility databases and the structural effects identified thereon could provide useful guidance for IL selection in related applications.