Hybrid data-driven and physics-based modeling for viscosity prediction of ionic liquids

Viscosity is one of the most important fundamental properties of fluids.However,accurate acquisition of viscosity for ionic liquids(ILs)remains a critical challenge.In this study,an approach integrating prior physical knowledge into the machine learning(ML)model was proposed to predict the viscosity reliably.The method was based on 16 quantum chemical descriptors determined from the first principles calculations and used as the input of the ML models to represent the size,structure,and interactions of the ILs.Three strategies based on the residuals of the COSMO-RS model were created as the output of ML,where the strategy directly using experimental data was also studied for comparison.The performance of six ML algorithms was compared in all strategies,and the CatBoost model was identified as the optimal one.The strategies employing the relative deviations were superior to that using the absolute deviation,and the relative ratio revealed the systematic prediction error of the COSMO-RS model.The CatBoost model based on the relative ratio achieved the highest prediction accuracy on the test set(R^(2)=0.9999,MAE=0.0325),reducing the average absolute relative deviation(AARD)in modeling from 52.45% to 1.54%.Features importance analysis indicated the average energy correction,solvation-free energy,and polarity moment were the key influencing the systematic deviation.

Melting points of ionic liquids:Review and evaluation

The melting points of ionic liquids(ILs)reported since 2020 were surveyed,collected,and reviewed,which were further combined with the previous data to provide a database with 3129 ILs ranging from 177.15 to 645.9 K in melting points.In addition,the factors that affect the melting point of ILs from macro,micro,and thermodynamic perspectives were summarized and analyzed.Then the development of the quantitative structure-property relationship(QSPR),group contribution method(GCM),and conductor-like screening model for realistic solvents(COSMO-RS)for predicting the melting points of ILs were reviewed and further analyzed.Combined with the evaluation together with the preliminary study conducted in this work,it shows that COSMO-RS is more promising and possible to further improve its performance,and a framework was thus proposed.

Silicon-based dielectric elastomer with amino-complexed hybrids towards high actuation performance

For improving the actuation performance at low electric fields of dielectric elastomers,achieving high dielectric constant(εr)and low modulus(Y)simultaneously has been targeted in the past decades,but there are few ways to accomplish both.In contrast to the classical strategies such as incorporating plasticizers or ceramic to prepare the silicon-based dielectric elastomers,here,blending an amino-complexed hybrid(polyethyleneimine(PEI)-Ag)with polydimethylsiloxane(PDMS)elastomer is reported as an alternative strategy to tailor theεr and Y.PEI-Ag not only exhibits excellent dielectric enhancement properties but also minimizes the PDMS crosslinking through amino-complexed reaction between PEI and Pt catalysts.The prepared dielectric elastomers have aεr of 7.2@10^(3)Hz and Y of 1.14 MPa,leading to an actuation strain of 22.27%at 35 V/μm.Hence,incorporating such novel hybrids based on dual amino-complexed effect on both matrix and particles sufficiently promotes the actuated performance of dielectric elastomers.

Ionic liquids/deep eutectic solvents for CO_(2) capture:Reviewing and evaluating

The CO_(2)solubilities(including CO_(2)Henry’s constant)in physical-and chemical-based ILs/DESs and the COSMO-RS models describing these properties were comprehensively collected and summarized.The summarized results indicate that chemical-based ILs/DESs are superior to physical-based ILs/DESs for CO_(2)capture,especially those ILs have functionalized cation and anion,and superbase DESs;some of the superbase DESs have higher CO_(2)solubilities than those of ILs;the best physical-and chemical-based ILs,as well as physical-and chemicalbased DESs are[BMIM][BF4](4.20 mol kg^(-1)),[DETAH][Im](11.91 mol kg^(-1)),[L-Arg]-Gly 1:6(4.92 mol kg^(-1))and TBD-EG 1:4(12.90 mol kg^(-1)),respectively.Besides the original COSMO-RS mainly providing qualitative predictions,six corrected COSMO-RS models have been proposed to improve the prediction performance based on the experimental data,but only one model is with universal parameters.The newly determined experimental results were further used to verify the perditions of original and corrected COSMO-RS models.The comparison indicates that the original COSMO-RS qualitatively predicts CO_(2)solubility for some but not all ILs/DESs,while the quantitative prediction is incapable at all.The original COSMO-RS is capable to predict CO_(2)Henry’s constant qualitatively for both physical-based ILs and DESs,and quantitative prediction is only available for DESs.For the corrected COSMO-RS models,only the model with universal parameters provides quantitative predictions for CO_(2)solubility in physical-based DESs,while other corrected models always show large deviations(>83%)compared with the experimental CO_(2)Henry’s constants.

Prediction and verification of heat capacities for pure ionic liquids

The heat capacity of ionic liquids is an important physical property,and experimental measuring is usually used as a common method to obtain them.Owing to the huge number of ionic liquids that can be potentially synthesized,it is desirable to acquire theoretical predictions.In this work,the Conductor-like Screening Model for Real Solvents(COSMO-RS)was used to predict the heat capacity of pure ionic liquids,and an intensive literature survey was conducted for providing a database to verify the prediction of COSMO-RS.The survey shows that the heat capacity is available for 117 ionic liquids at temperatures ranging 77.66-520 K since 2004,and the 4025 data points in total with the values from 76.37 to 1484 J·mol^(-1)·K^(-1) have been reported.The prediction of heat capacity with COSMO-RS can only be conducted at two temperatures(298 and 323 K).The comparison with the experimental data proves the prediction reliability of COSMO-RS,and the average relative deviation(ARD)is 8.54%.Based on the predictions at two temperatures,a linear equation was obtained for each ionic liquid,and the heat capacities at other temperatures were then estimated via interpolation and extrapolation.The acquired heat capacities at other temperatures were then compared with the experimental data,and the ARD is only 9.50%.This evidences that the heat capacity of a pure ionic liquid follows a linear equation within the temperature range of study,and COSMO-RS can be used to predict the heat capacity of ionic liquids reliably.