COSMO‐UNIFAC model for ionic liquids

Dong, Yichun; Huang, Shuai; Guo, Yali; Lei, Zhigang

doi:10.1002/aic.16787

Cited by 34 publications

(36 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, the solute type specific comparison is also made between the RS model predictions with the UNIFAC‐IL calculations, where the lower MAPE of the RS model is generally observed with only an exception for the carboxylic acids. This finding also holds true when comparing the RS model with the UNIFAC and COSMO‐UNIFAC models reported by Dong et al, 21 although only a small number of experimental data are available for the mutual comparison (see detailed data and statistics in Table S3). To summarize, the developed RS model quantitatively outperforms the state‐of‐the‐art methods for predicting solute‐in‐IL γ ∞ .…”

Section: Dnn‐based Rs For γ∞ Predictionsupporting

confidence: 72%

“…It can be seen that all the binary interaction parameters involved in the IL‐solute group matrix are filled in the UNIFAC‐IL extension thanks to the advantage of the RS established above. Again, the obtained UNIFAC‐IL model is compared with the UNIFAC and COSMO‐UNIFAC reported by Dong et al 21 As tabulated in Table S7, for the seven involved binary systems with overall 412 data points (Acetonitrile + [BMIM][BF 4 ], Pyridine + [EMIM][SCN], Acetone + [EMIM][CF 3 SO 3 ], Thiophene + [BMIM][PF 6 ], Methanol + [EMIM][DEP], Acetone + [EMIM][DEP], Ethanol + [EMIM][DEP]), the MAPEs between experimental γ and predictions by different methods follows the ranking of: COSMO‐UNIFAC (21.50%) > UNIFAC‐IL (16.79%) ≈ UNIFAC (16.54%). The UNIFAC model by Dong et al presents the lowest MAPE here as this model is essentially regressed from these VLE data.…”

Section: Application Of Rsmentioning

confidence: 96%

“…The very recently reported UNIFAC‐IL model based on experimental γ ∞ is selected here as it has already covered 619 pairs of binary interaction parameters, which was proved with good γ ∞ prediction reliability and also encouraging extrapolation performance for liquid–liquid and vapor–liquid equilibria prediction 16 . Of course, one can also employ the completed γ ∞ matrix for the extension of other versions of UNIFAC‐IL model by other researchers 20,21 …”

Section: Application Of Rsmentioning

confidence: 99%

“…Due to the fully predictive character, the COSMO‐based activity coefficient methods are virtually applicable to arbitrary pure liquid and liquid mixtures, including ILs and IL‐involving mixtures; but on the flip side, predictions by COSMO‐based methods are, in some cases, only qualitative rather than quantitative. Taking account of merits of the UNIFAC and COSMO‐based methods, the usage of COSMO‐based activity coefficient predictions for the regression of missing UNIFAC model parameters is also proposed by several researchers 16,19‐21 . This manner theoretically allows the extension of the applicability range of UNIFAC to any mixture of interest.…”

Section: Introductionmentioning

confidence: 99%

“…Taking account of merits of the UNIFAC and COSMO-based methods, the usage of COSMO-based activity coefficient predictions for the regression of missing UNIFAC model parameters is also proposed by several researchers. 16,[19][20][21] This manner theoretically allows the extension of the applicability range of UNIFAC to any mixture of interest.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Neural recommender system for the activity coefficient prediction andUNIFACmodel extension of ionicliquid‐solutesystems

Chen

Song

et al. 2021

AIChE Journal

View full text Add to dashboard Cite

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∞ prediction 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.

show abstract

Section: Dnn‐based Rs For γ∞ Predictionsupporting

confidence: 72%

Section: Application Of Rsmentioning

confidence: 96%

Section: Application Of Rsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Neural recommender system for the activity coefficient prediction andUNIFACmodel extension of ionicliquid‐solutesystems

Chen

Song

et al. 2021

AIChE Journal

View full text Add to dashboard Cite

show abstract

Predictive molecular thermodynamic models for ionic liquids

Wei

Chen

et al. 2022

AIChE Journal

Self Cite

View full text Add to dashboard Cite

Predictive molecular thermodynamic models can bridge the gap between the microscopic molecular level and macroscopic system scale. Over the past few decades, ionic liquids (ILs), as a class of green solvents and functional materials, have received widespread research interest in the field of chemical processing owing to their unique physicochemical merits. This review aims to provide an easy‐to‐read and exhaustive reference regarding state‐of‐the‐art predictive thermodynamic models for ILs, with more focuses on UNIFAC‐ and COSMO‐based models and various applications involving phase equilibrium prediction, guidance for IL screening and design, and building equilibrium stage models for separation processes. This review provides a theoretical perspective on the structure–property relationships between molecular structures and phase behaviors for the systems and the constituted components covering a multi‐scale viewpoint from molecular level to industrial scale. Moreover, the predictive capacities of different thermodynamic models are comprehensively compared.

show abstract

Solutions, in Particular Dilute Solutions of Nonelectrolytes: A Review

Wilhelm¹

2021

J Solution Chem

View full text Add to dashboard Cite

The liquid state is one of the three principal states of matter and arguably the most important one; and liquid mixtures represent a large research field of profound theoretical and practical interest. This topic is of importance in many areas of the applied sciences, such as in chemical engineering, geochemistry, the environmental sciences, biophysics and biomedical technology. First, I will concisely present a review of important concepts from classical thermodynamics of nonelectrolyte solutions; this will be followed by a survey of (semi-)empirical approaches to representing the composition and temperature dependence of selected thermodynamic mixture properties, and finally the focus will be on dilute binary nonelectrolyte solutions where one component, a supercritical solute, is present in much smaller quantity than the other component, called the solvent. Partial molar properties in the limit of infinite dilution (indicated by a superscript ∞) are of particular interest. For instance, activity coefficients (Lewis–Randall (LR) convention) are customarily used to characterize mixing behavior, and infinite-dilution values $$\gamma_{i}^{{{\text{LR,}}\infty }}$$ γ i LR, ∞ provide a convenient route for obtaining binary parameters for several popular solution models. When discussing solute (j)—solvent (i) interactions in solutions where the solute is supercritical, the Henry fugacity $$h_{j,i} \left( {T,P} \right)$$ h j , i T , P , also known as Henry’s law (HL) constant, is a measurable thermodynamic key quantity. Its temperature dependence yields information on the partial molar enthalpy change on solution $$\Delta H_{j}^{\infty } \left( {T,P} \right)$$ Δ H j ∞ T , P , while its pressure dependence yields information on the partial molar volume $$V_{j}^{{{\text{L,}}\infty }} \left( {T,P} \right)$$ V j L, ∞ T , P of solute j in the liquid phase (superscript L). I will clarify issues frequently overlooked, touch upon solubility data reduction and correlation, report a few recent high-precision experimental results on dilute aqueous solutions of supercritical nonelectrolytes, and show the equivalency of results for caloric quantities (e.g. $$\Delta H_{j}^{\infty }$$ Δ H j ∞ ) obtained via van ’t Hoff analysis of high-precision solubility data with directly measured calorimetric data.

show abstract

COSMO‐UNIFAC model for ionic liquids

Cited by 34 publications

References 12 publications

Neural recommender system for the activity coefficient prediction andUNIFACmodel extension of ionicliquid‐solutesystems

Neural recommender system for the activity coefficient prediction andUNIFACmodel extension of ionicliquid‐solutesystems

Predictive molecular thermodynamic models for ionic liquids

Solutions, in Particular Dilute Solutions of Nonelectrolytes: A Review

Contact Info

Product

Resources

About