Of central importance to evaluate the suitability of ionic liquids (ILs) for a process is the accurate estimation of IL properties related to target performances. In this work, a versatile deep learning method for predicting IL properties is developed. Molecular fingerprints are derived from the encoder state of a Transformer model pre-trained on the PubChem database, which allows transfer learning from large-scale unlabeled data and significantly improves generalization performance for developing models with small datasets. Employing the pre-trained molecular fingerprints, convolutional neural network (CNN) models for IL properties prediction are trained and tested on 11 databases. The obtained Transformer-CNN models present superior performance to state-of-the-art models in all cases and enable property prediction of millions of ILs shortly. The application of the proposed models is exemplified by searching CO2 absorbent from a huge database of 8,333,096 synthetically feasible ILs, which is by far the most high-throughput IL screening in literature.

For the ionic liquid (IL)-solute systems of broad interest, a deep neural network based recommender system (RS) for predicting the infinite dilution activity coefficient (γ∞) is proposed and applied for a large extension of the UNIFAC model. In the RS, neural network entity embeddings are employed for mapping each IL and solute and neural collaborative filtering is utilized to handle the nonlinearities of IL-solute interactions. A comprehensive experimental γ∞ database covering 215 ILs and 112 solutes (totally 41,553 data points) is established for training the RS, where the cross-validation and test are performed based on a rigorous dataset split by IL-solute combinations. The obtained RS shows superior performance than the state-of-the-art γ∞ models and is thus taken to complete the solute-in-IL γ∞ matrix. Based on the completed γ∞ database, a large extension of the UNIFAC-IL model is finally presented, filling all the parameters between involved groups.