Search

Ionic liquids (ILs) as green solvents were first used to capture 1,2-dimethoxyethane (DMET) from exhaust gas. The vapor liquid equilibrium (VLE) data of [EMIM][TF2N] + DMET and triethylene glycol (TEG) + DMET were measured. The results showed that the popular UNIFAC-Lei model is in good agreement with experimental data. The gas absorption experiment with [EMIM][TF2N] as absorbent was conducted. It demonstrates that the IL has good separation performance for capturing DMET, the absorption ratio being 87.32%. Then, the absorption capacities of 45 ILs for DMET were calculated by COSMO-RS model, while DMET removal mechanism was explored by σ-profiles analysis and molecular dynamics (MD) simulation. Finally, industrial scale process simulation using [EMIM][TF2N] and TEG as absorbents to capture DMET was made. Compared to the benchmark TEG process, the total operating cost and the total annual cost (TAC) in the IL process decrease by 24.06% and 17.15%, respectively. This further proves that ILs have great application potential for the condensable volatile organic compounds (VOCs) capture, especially in energy saving and improving economic benefits.

Ionic liquids (ILs) as separating agents have been extensively studied in chemical separation processes. For IL-containing systems, the UNIFAC model for ILs (i.e., UNIFAC-Lei model) has become a popular group contribution approach to predict phase equilibrium behavior. Since the UNIFAC-Lei model was established in 2009, our group has continued to extend and revise this model because the accurate prediction and application scope of this model depend on the quality and size of the parameter matrix besides the statements and hypothesis assumed for the model. Thus, this work is focused on revising and extending the UNIFAC-Lei parameter matrix through plentiful and accurate experimental data. In this work, four new groups ([MIM][DBP], [MIM][TCM], [MIM][DCA], and [PYM][TCM]) were added, 63 pairs of new parameters were extended, and eight pairs of old parameters were revised. The COSMO calculation was used to get the unknown group surface area and volume parameters (Rkand Qk) of IL groups; the experimental phase equilibrium data of the systems containing ILs were used to get the unknown group binary interaction parameters (anmand amn). The comparison of the UNIFAC-Lei predicted vapor–liquid equilibrium values with experimental data indicates that the UNIFAC-Lei model can provide relatively satisfactory phase equilibrium prediction for IL-containing systems.

In this work, a prioriCOSMO-SAC model was combined with the original UNIFAC model, and thus the new COSMO-SAC-UNIFAC thermodynamic model with a strong predictive power for fluid phase equilibrium (VLE, LLE) was proposed. By this means the group binary parameter matrix of original UNIFAC model was extended by introducing the new 648 vacant parameter pairs for 51 main functional groups for the conventional substances in this work. Moreover, the combined thermodynamic model was first applied to process simulation on gas drying with ionic liquids (ILs). To verify the reliability of the COSMO-SAC-UNIFAC model, the predictive values were directly compared with experimental data coming from this work, our previous work, and literature. The moderately accurate predictions of the COSMO-SAC-UNIFAC model demonstrate the high potential applicability of this new model, especially for numerous systems with missing parameters including ILs, which are normally encountered in the original UNIFAC model.

Predictive molecular thermodynamic models are significant for predicting basic thermodynamic properties, such as fluid phase equilibrium in chemical separation processes and the subsequent simulation and amplification of industrial processes. Considering that some group binary interaction parameters between ionic liquid (IL) groups and traditional substance groups in the original UNIFAC model are still missing, 18 pairs of parameters are obtained by collecting experimental data from the new literature. On the other hand, the remaining 1570 pairs of interaction parameters are expanded by the COSMO-UNIFAC model, with other parameters missing from the literature data. The prediction of gas–liquid equilibrium, liquid–liquid equilibrium, gas–liquid equilibrium, and infinite dilution activity coefficient is used to evaluate the prediction performance of the newly extended UNIFAC model for phase equilibria of systems containing ionic liquids.

A united chemical thermodynamic model, that is, the COSMO-UNIFAC model, was first proposed to predict the phase equilibrium of multicomponent systems in which the UNIFAC model parameters are missing. This model combines the advantages of the UNIFAC model (accurate prediction) and the COSMO-based models (a priori prediction). The predicted vapor–liquid equilibrium results by the COSMO-UNIFAC model were compared with experimental data from the literature and this work, confirming that it can provide a moderate quantitative prediction even if the UNIFAC model parameters are missing.

Ionic liquids (ILs) have been proposed to simultaneously capture a variety of condensable gases for the first time. The hydrophobic IL [EMIM][Tf2N] was selected as a suitable absorbent screened by the COSMO-RS (conductor-like screening model for real solvents) model, which in combination with the quantum chemistry calculation provided some theoretical insight into mechanisms with respect to condensable gas capture. Considering the simultaneous capture of volatile organic compounds (VOCs) and water, we selected benzene, toluene, and p-xylene (BTX) as three kinds of VOC representatives, measured the vapor–liquid equilibrium (VLE) of BTX + [EMIM][Tf2N] mixture systems, and compared them with the predicted results by the UNIFAC–Lei model. The experiment of capturing condensable gases with [EMIM][Tf2N] as the absorbent was conducted. Furthermore, the conceptual design of continuous processes with [EMIM][Tf2N] and the conventional benchmark solvent triethylene glycol as absorbents at the industrial scale was performed using a rigorous equilibrium stage model with the UNIFAC–Lei model parameter input. It reveals that condensable gas capture with an IL belongs to a typical process intensification technology in chemical engineering.

A suitable ionic liquid for methyl chloride drying experiment was screened out from 210 ionic liquids by COSMO-RS model. Moreover, the experimental mechanism of ionic liquids drying is further explained by the COSMO-RS model, and it is further confirmed by analyzing the binding energy. The solubility of methyl chloride in [EMIM][BF4] and TEG and [EMIM][BF4]+H2O was completed, and the experimental results well proved the reliability of the UNIFAC-Lei model. The unknown interaction parameters were obtained through the solubility data of this work and the experimental data in the literatures. The methyl chloride drying experiment was completed in the laboratory, and the water content of the methyl chloride can be reduced to below 200 ppm. The simulation of the methyl chloride drying process using [EMIM][BF4] or TEG as absorbents was carried out by ASPEN software on an industrial scale. The final simulation results show that the [EMIM][BF4] drying process has lower energy consumption and better drying effect.

The ionic liquid (IL) 1-ethyl-3-methylimidazolium dicyanamide ([EMIM][DCA]) was selected as an appropriate entrainer for the extractive distillation of the methanol–ethanol-water mixture. The COSMO-RS model was applied to screen out the appropriate solvents considering selectivity and solvent capacity together. Isobaric vapor–liquid equilibrium (VLE) experiments for the two systems of methanol–water and ethanol–water with different amounts of [EMIM][DCA] added were conducted at 101.3 kPa. The experimental data showed that [EMIM][DCA] exhibits an obvious salting effect for the methanol (or ethanol)-water mixture and eliminates the azeotropic point of ethanol–water. Moreover, the predicted values by UNIFAC-Lei model coincide well with experimental data. The separation mechanism was further explained in combination with surface charge density distribution (σ-profiles), excess enthalpy (HE), and binding energy. In addition, the flow charts were designed to evaluate the improvement of energy consumption with [EMIM][DCA] as the entrainer when compared to ethylene glycol (EG). The simulation results demonstrated that [EMIM][DCA] is more energy efficient than EG.

We like to highlight the extension of COSMO-based models (COSMO RS-FV and COSMO SAC-FV) to the prediction of gas solubility in polymers, including polymerized ionic liquids (PILs) for the first time. To verify the applicability of COSMO-based models, the predicted values for gas solubility in both common polymers (CH4/N2 + PEG) and PILs (CO2 + P[MATMA][BF4]/P[VATMA][BF4]) were evaluated based on previous experimental data. It was confirmed that the COSMO-RS (Klamt) model performs better than the COSMO-SAC model for common polymers, whereas the COSMO-RS (ADF-Lei 2018) exhibits the best predictions for PILs. The moderately accurate predictions of COSMO-based models demonstrate the high potential for predicting gas solubility in polymers.