Viscosity model of deep eutectic solvents from group contribution method

Abstract

Deep eutectic solvents (DESs), a novel category of sustainable solvents, are expected to achieve the design of the chemical processes without utilizing or generating harmful chemicals. In this work, based on the mathematical model inspired by the transition state theory, the group contribution method is used to accurately predict the viscosity of DESs. The model is constrained by Eyring rate theory and hard sphere free volume theory. A dataset of 2229 experimental viscosity data points of 183 DESs from literature is used to determine the model parameters and subsequently verify the model. The rules introduced by this model are simple and easy to follow. The results show that the proposed model is capable to predict the viscosity of DESs with very high accuracy, using only temperature and composition as inputs. The average absolute relative deviations (AARDs) of the model are 8.12% and 8.64% over the training and test sets, respectively, and the maximum ARD is 34.63%. Therefore, the as-proposed model can be considered a highly reliable tool for predicting the viscosity of DESs when experimental data are absent. It will provide useful guidance for the synthesis of DESs with specific viscosity to meet different application requirements and promote their industrial-scale implementation.

Modeling the Viscosity of ChCl-Based Deep Eutectic Solvents and Their Mixtures with Water

“… a AARD = 1 N false ∑ i = 1 N true| η i , exp ‐ η i , pre true| η i , exp ; MAE = 1 N false ∑ i = 1 N true| η i , exp ‐ η i , pre true| ; the underlined ChCl-based DESs are not included in the dataset to develop the GC and ML models; dash line means the AARD is higher than 100%. …”

Design Strategy and Application of Deep Eutectic Solvents for Green Synthesis of Nanomaterials

“…It was observed that DESs with bromide anions are more viscous than chloride anions and lower viscosities DESs only contain long-chain fatty acids. The viscosity model of DESs obtained from the group contribution method has been recently reported [ 42 ]. The model predicted high accuracy of DESs viscosity using input parameters of temperature and composition and could be a powerful tool for discovering the viscosity of DESs.…”

Deep eutectic solvents as a versatile platform toward CO2 capture and utilization

“…Moreover, the viscosity prediction of DESs has also been explored using quantitative structure–property relationship (QSPR) models and their combination with machine learning methods. 55,56 The group contribution (GC) model, which combines Eyring rate and free volume theory, has been developed to evaluate temperature-dependent viscosity of DESs. 56 For 2229 collected experimental viscosity data points related to 183 DESs, the average absolute relative deviations of the GC model are 8.12% and 8.64% over training and test sets (1486 and 743 data points), respectively.…”

“…55,56 The group contribution (GC) model, which combines Eyring rate and free volume theory, has been developed to evaluate temperature-dependent viscosity of DESs. 56 For 2229 collected experimental viscosity data points related to 183 DESs, the average absolute relative deviations of the GC model are 8.12% and 8.64% over training and test sets (1486 and 743 data points), respectively. Therefore, the viscosity prediction models of DESs should be reliable even in the absence of experimental data.…”