A carregar...
Projeto de investigação
Developing extensive thermodynamic and machine learning models to predict various physical properties and phase equilibria of deep eutectic solvents and incorporation into a comprehensive commercial software
Financiador
Autores
Publicações
Application of Hansen solubility parameters in the eutectic mixtures
Publication . Fernandes, Cláudio C.; Paiva, Alexandre; Haghbakhsh, Reza; Duarte, Ana Rita C.; DQ - Departamento de Química; LAQV@REQUIMTE; Nature Publishing Group
Hansen Solubility Parameters (HSPs) are widely used as a tool in solubility studies. Given the variety of existent approaches to predict these parameters, this investigation focused on estimating the HSPs of a set of Natural Deep Eutectic Systems (NADES), using empirical (EM) and semi-empirical models (SEM), and then understanding their differences/similarities. Although these theoretical models are designed and recommended mostly for simple molecules or simple solutions, they are still being used in eutectic systems studies, mainly empirical ones. Thus, a preliminary test was conducted with a set of conventional solvents, in which their experimental values of HSPs are known. Besides the confirmation of the EM as the most suitable for these kinds of regular solvents, the results found also showed a very similar behaviour to what was observed in NADES, i.e., in terms of suggesting the EM and SEM with the highest/lowest similarity. Furthermore, it was concluded that although there is a large discrepancy between the estimated values of the hydrogen bond parameter, especially for systems with a higher polar character, there is still a good similarity for the other parameters. In fact, it was observed that, when combining the semi-empirical models, it was possible to obtain a value of the hydrogen bond parameter more similar to the empirical ones.
Understanding the Solubility behaviour of Ibuprofen and Xylitol in Natural Deep Eutectic Systems through Hansen Solubility Parameters and Physicochemical Properties
Publication . Fernandes, Cláudio C.; Paiva, Alexandre; Haghbakhsh, Reza; Duarte, Ana Rita C.; LAQV@REQUIMTE; Elsevier
Most of the recent studies have been praising the peculiar ability of deep eutectic systems (DES), especially the natural-based ones (designated by NADES), in dissolving a wide variety of compounds. Despite their remarkable physicochemical properties, it is still true that little is known about the factors that would help to comprehend their interesting behaviour when they are used as solvents. Hence, it is important to gather as many tools as possible that can be useful for understanding it. First, the affinity degrees between the two selected compounds (Ibuprofen and Xylitol) and the various NADES, were analysed using Hansen Solubility Parameters (HSPs), which confirmed to be a good tool for screening good and bad NADES for solubilising Ibuprofen and Xylitol. Although, in general, the empirical models (EM) such as the one proposed by Hoftyzer-Van Krevelen and Fedors (HKF) and Yamamoto (Ymt) performed better than the semi-empirical models (SEM), when it came to assessing affinity, it was found that this actually depends on the type of assessment carried out, i.e., if it is in 1- or 2-dimension. Furthermore, it was also found that, except for the dispersive parameter (δd), all the others play a significant role in the interaction between the two compounds and NADES, especially the total solubility parameter (δt). Finally, the correlations between a set of physiochemical properties of NADES and the solubility data were evaluated in this work where it was possible to conclude that surface tension, density and molar volume are those that present the highest contribution for the variations in the solubility.
An insight into the CO2 solubility in NaCl + ethylene glycol (1:16) deep eutectic solvent
Publication . Rasoolzadeh, Ali; Keshtkar, Mehdi; Raeissi, Sona; Haghbakhsh, Reza; LAQV@REQUIMTE; DQ - Departamento de Química; Elsevier BV
The regulation of CO₂ emissions from industrial operations is crucial from an environmental perspective. The most widely used solvents for CO₂ capture consist of aqueous alkanolamine solutions. However, amine-based processes face several challenges, such as corrosion, chemical degradation, and high energy requirements for solvent regeneration. As potential alternatives, deep eutectic solvents (DESs) have emerged as promising eco-friendly and biodegradable options for CO₂ capture. This study experimentally measures the solubility of CO₂ in a DES (1 mol NaCl + 16 mol ethylene glycol) using a high-pressure solubility apparatus at the four temperatures of 293.15, 303.15, 313.15, and 323.15 K. For the thermodynamic modeling, the Soave Redlich Kwong equation of state (SRK EoS) was employed, coupled with three different mixing rules of van der Waals (vdW), Wong Sandler (WS), and modified Huron-Vidal (MHV1). The vdW approach was considered in the three cases without using binary interaction, constant binary interaction, and variable binary interaction parameter by temperature. The results demonstrated that by incorporating the WS and MHV1, the local composition concept was successful in addressing the non-ideality of the liquid phase. Among the tested models, the WS (AARD%=6.66) and MHV1 (AARD%=5.61) provided the most accurate predictions of equilibrium pressures. Additionally, Henry's constant, standard Gibbs energy, enthalpy, and entropy of gas solvation were determined using the experimental data together with classical thermodynamic relations. The calculated negative standard enthalpy of solvation indicates an exothermic gas solvation process, signifying that energy is released as CO₂ dissolves in this DES.
Machine learning-driven prediction of deep eutectic solvents’ heat capacity for sustainable process design
Publication . Halder, Amit Kumar; Haghbakhsh, Reza; Ferreira, Elisabete S. C.; Duarte, Ana Rita C.; Cordeiro, M. Natália D. S.; LAQV@REQUIMTE; DQ - Departamento de Química; Elsevier
Heat capacity, a crucial physical property for chemical processes, is often understudied in Deep Eutectic Solvents (DESs), which in turn are promising green alternatives to environmentally hazardous conventional solvents. This work addresses this gap by developing a machine learning model to predict DES heat capacity and identify key structural features influencing it. We employed a dataset of 530 DESs with corresponding experimental heat capacity values. Quantum-chemical COSMO-RS-based descriptors, capturing detailed information about DES structures, were calculated for each data point. Various machine learning algorithms, namely k-Nearest Neighbours (kNN), Random Forests (RF), Neural Network Multilayer Perceptron (MLP), and Support Vector Machines (SVM) were explored alongside a linear model (Multiple Linear Regression, MLR). Hyperparameter optimisation ensured all models were fine-tuned for optimal performance. The most successful model, based on the MLP technique, achieved remarkably low Average Absolute Relative Deviation (AARD) values of 0.500 % and 3.999 % for the training and test sets, respectively. This signifies a significant improvement in prediction accuracy compared to traditional methods. Furthermore, by applying a SHapley Additive exPlanations (SHAP) analysis, we identified the most crucial structural factors within DES components that govern their heat capacity. This comprehensive investigation offers valuable insights that can pave the way for an efficient design of novel DESs in the future.
Global atomic and group contribution models for prediction of the thermal conductivities of deep eutectic solvents
Publication . Soltani, Fatemeh; Haghbakhsh, Reza; Raeissi, Sona; LAQV@REQUIMTE; DQ - Departamento de Química; Elsevier Science B.V., Amsterdam.
Deep Eutectic Solvents (DESs) are often categorized as novel green solvents. Knowledge of the thermal conductivity of a solvent in an industrial process is vital for the optimization of energy utilization. Considering the vast number of DESs introduced to date, it is practically impossible to measure all their thermal conductivities. Thus, it is vital to have predictive models that can predict the thermal conductivities of various DESs, and at different temperatures. For this purpose, a large data bank was collected, including 338 data points from 56 DESs of various natures. The data were used to develop a group contribution (GC) model and an atomic contribution (AC) model to predict the thermal conductivities of DESs. The calculated AARD% values of 7.62 % and 9.52 % for the proposed GC and AC models, respectively, indicated reliable performance and promising predictions for both models. The models were also compared to well-known literature models.
Unidades organizacionais
Descrição
Palavras-chave
Contribuidores
Financiadores
Entidade financiadora
Fundação para a Ciência e a Tecnologia
Programa de financiamento
CEEC IND5ed
Número da atribuição
2022.05803.CEECIND/CP1725/CT0003