Deep eutectic solvents (DESs) are one of the most rapidly evolving types of solvents, appearing in a broad range of applications, such as nanotechnology, electrochemistry, biomass transformation, pharmaceuticals, membrane technology, biocomposite development, modern 3D-printing, and many others. The range of their applicability continues to expand, which demands the development of new DESs with improved properties. To do so requires an understanding of the fundamental relationship between the structure and properties of DESs. Computer simulation and machine learning techniques provide a fruitful approach as they can predict and reveal physical mechanisms and readily be linked to experiments. This review is devoted to the computational research of DESs and describes technical features of DES simulations and the corresponding perspectives on various DES applications. The aim is to demonstrate the current frontiers of computational research of DESs and discuss future perspectives.

Due to the great importance for many industrial applications it is crucial from the point of view of theoretical description to reproduce thermal properties of thermoplastic polyimides as accurate as possible in order to establish “chemical structure–physical properties” relationships of new materials. In this paper we employ differential scanning calorimetry, dilatometry, and atomistic molecular dynamics (MD) simulations to explore whether the state-of-the-art computer modeling can serve as a precise tool for probing thermal properties of polyimides with highly polar groups. For this purpose the polyimide R–BAPS based on dianhydride 1,3-bis(3′,4-dicarboxyphenoxy)benzene (dianhydride R) and diamine 4,4′-bis(4′′-aminophenoxy)biphenyl sulphone) (diamine BAPS) was synthesized and extensively studied. Overall, our findings show that the widely used glass-transition temperature T_g evaluated from MD simulations should be employed with great caution for verification of the polyimide computational models against experimental data: in addition to the well-known impact of the cooling rate on the glass-transition temperature, correct definition of T_g requires cooling that starts from very high temperatures (no less than 800 K for considered polyimides) and accurate evaluation of the appropriate cooling rate, otherwise the errors in the measured values of T_g become undefined. In contrast to the glass-transition temperature, the volumetric coefficient of thermal expansion (CTE) does not depend on the cooling rate in the low-temperature domain (T < T_g) so that comparison of computational and experimental values of CTE provides a much safer way for proper validation of the theoretical model when electrostatic interactions are taken into account explicitly. Remarkably, this conclusion is most likely of generic nature: we show that it also holds for the commercial polyimide EXTEM™, another polyimide with a similar chemical structure.

The results of atomistic molecular-dynamics simulations of mechanical properties of heterocyclic polymer subjected to uniaxial deformation are reported. A new amorphous thermoplastic polyimide R–BAPO with a repeat unit consisting of dianhydride 1,3-bis-(3′,4,-dicarboxyphenoxy)diphenyl (dianhydride R) and diamine 4,4′-bis-(4′′-aminophenoxy)diphenyloxide (diamine BAPO) was chosen for the simulations. Our primary goal was to establish the impact of various factors (sample preparation method, molecular mass, and cooling and deformation rates) on the elasticity modulus. In particular, we found that the elasticity modulus was only slightly affected by the degree of equilibration, the molecular mass and the size of the simulation box. This is most likely due to the fact that the main contribution to the elasticity modulus is from processes on scales smaller than the entanglement length. Essentially, our simulations reproduce the logarithmic dependence of the elasticity modulus on cooling and deformation rates, which is normally observed in experiments. With the use of the temperature dependence analysis of the elasticity modulus we determined the flow temperature of R–BAPO to be 580 K in line with the experimental data available. Furthermore, we found that the application of high external pressure to the polymer sample during uniaxial deformation can improve the mechanical properties of the polyimide. Overall, the results of our simulations clearly demonstrate that atomistic molecular-dynamics simulations represent a powerful and accurate tool for studying the mechanical properties of heterocyclic polymers and can therefore be useful for the virtual design of new materials, thereby supporting cost-effective synthesis and experimental research.

In the present work, we address the problem of utilizing machine learning (ML) methods to predict the thermal properties of polymers by establishing "structure-property" relationships. Having focused on a particular class of heterocyclic polymers, namely polyimides (PIs), we developed a graph convolutional neural network (GCNN), being one of the most promising tools for working with big data, to predict the PI glass transition temperature T _g as an example of the fundamental property of polymers. To train the GCNN, we propose an original methodology based on using a "transfer learning" approach with an enormous "synthetic" data set for pretraining and a small experimental data set for its fine-tuning. The "synthetic" data set contains more than 6 million combinatorically generated repeating units of PIs and theoretical values of their T _g values calculated using the well-established Askadskii's quantitative structure-property relationship (QSPR) computational scheme. Additionally, an experimental data set for 214 PIs was also collected from the literature for training, fine-tuning, and validation of the GCNN. Both "synthetic" and experimental data sets are included into a PolyAskInG database (Polymer Askadskii's Intelligent Gateway). By using the PolyAskInG database, we developed GCNN which allows estimation of T _g of PI with a mean absolute error (MAE) of about 20 K, which is 1.5 times lower than in the case of Askadskii QSPR analysis (33 K). To prove the efficiency and usability of the proposed GCNN architecture and training methodology for predicting polymer properties, we also employed "transfer learning" to develop alternative GCNN pretrained on proxy-characteristics taken from the popular quantum-chemical QM9 database for small compounds and fine-tuned on an experimental T _g values data set from PolyAskInG database. The obtained results indicate that pretraining of GCNN on the "synthetic" polymer data set provides MAE which is almost twice as low as that in the case of using the QM9 data set in the pretraining stage (∼41 K). Furthermore, we address the questions associated with the influence of the differences in the size of the experimental and "synthetic" data sets (so-called "reality gap" problem), as well as their chemical composition on the training quality. Our results state the overall priority of using polymer data sets for developing deep neural networks, and GCNN in particular, for efficient prediction of polymer properties. Moreover, our work opens up a challenge for the theoretically supported generation of large "synthetic" data sets of polymer properties for the training of the complex ML models. The proposed methodology is rather versatile and may be generalized for predicting other properties of different polymers and copolymers synthesized through the polycondensation reaction.