Explainable Solvation Free Energy Prediction Combining Graph Neural Networks with Chemical Intuition.

The prediction of a molecule's solvation Gibbs free (ΔGsolv) energy in a given solvent is an important task which has traditionally been carried out via quantum chemical continuum methods or force field-based molecular simulations. Machine learning (ML) and graph neural networks in particular have emerged as powerful techniques for elucidating structure-property relationships. This work presents a graph neural network (GNN) for the prediction of ΔGsolv which, in addition to encoding typical atom and bond-level features, incorporates chemically intuitive, solvation-relevant parameters into the featurization process: semiempirical partial atomic charges and solvent dielectric constant. Solute-solvent interactions are included via an interaction map layer which can be visualized to examine solubility-enhancing or -decreasing interactions learnt by the model. On a test set of small organic molecules, our GNN predicts ΔGsolv in water and cyclohexane with an accuracy comparable to polarizable and ab initio generated force field methods [mean absolute error (MAE) = 0.4 and 0.2 kcal mol-1, respectively], without the need for any molecular simulation. For the FreeSolv data set of hydration free energies, the test MAE is 0.7 kcal mol-1. Interpretability and applicability of the model is highlighted through several examples including rationalizing the increased solubility of modified diaminoanthraquinones in organic solvents. The clear explanations afforded by our GNN allow for easy understanding of the model's predictions, giving the experimental chemist confidence in employing ML models toward more optimized synthetic routes.

What's in an Atom? A Comparison of Carbon and Silicon-Centred Amidinium⋠⋠⋠Carboxylate Frameworks*.

Despite their apparent similarity, framework materials based on tetraphenylmethane and tetraphenylsilane building blocks often have quite different structures and topologies. Herein, we describe a new silicon tetraamidinium compound and use it to prepare crystalline hydrogen bonded frameworks with carboxylate anions in water. The silicon-containing frameworks are compared with those prepared from the analogous carbon tetraamidinium: when biphenyldicarboxylate or tetrakis(4-carboxyphenyl)methane anions were used similar channel-containing networks are observed for both the silicon and carbon tetraamidinium. When terephthalate or bicarbonate anions were used, different products form. Insights into possible reasons for the different products are provided by a survey of the Cambridge Structural Database and quantum chemical calculations, both of which indicate that, contrary to expectations, tetraphenylsilane derivatives have less geometrical flexibility than tetraphenylmethane derivatives, that is, they are less able to distort away from ideal tetrahedral bond angles.

The effect of descriptor choice in machine learning models for ionic liquid melting point prediction.

The characterization of an ionic liquid's properties based on structural information is a longstanding goal of computational chemistry, which has received much focus from ab initio and molecular dynamics calculations. This work examines kernel ridge regression models built from an experimental dataset of 2212 ionic liquid melting points consisting of diverse ion types. Structural descriptors, which have been shown to predict quantum mechanical properties of small neutral molecules within chemical accuracy, benefit from the addition of first-principles data related to the target property (molecular orbital energy, charge density profile, and interaction energy based on the geometry of a single ion pair) when predicting the melting point of ionic liquids. Out of the two chosen structural descriptors, ECFP4 circular fingerprints and the Coulomb matrix, the addition of molecular orbital energies and all quantum mechanical data to each descriptor, respectively, increases the accuracy of surrogate models for melting point prediction compared to using the structural descriptors alone. The best model, based on ECFP4 and molecular orbital energies, predicts ionic liquid melting points with an average mean absolute error of 29 K and, unlike group contribution methods, which have achieved similar results, is applicable to any type of ionic liquid.

Accurate Prediction of Three-Body Intermolecular Interactions via Electron Deformation Density-Based Machine Learning.

This work extends the electron deformation density-based descriptor, originally developed in the electron deformation density-based interaction energy machine learning (EDDIE-ML) algorithm to predict dimer interaction energies, to the prediction of three-body interactions in trimers. Using a sequential learning process to select the training data, the resulting Gaussian process regression (GPR) model predicts the three-body interaction energy within 0.2 kcal mol-1 of the SRS-MP2/cc-pVTZ reference values for the 3B69 and S22-3 trimer data sets. A hybrid kernel function is introduced, which combines contributions from the average and individual atomic environments, allowing the total trimer interaction energy to be predicted in addition to the three-body contribution using the same descriptor. To extend the range and diversity of trimer interaction energies available in the literature, a new data set based on a protein-ligand crystal structure is introduced, consisting of 509 structures of a central ligand with two protein fragments. Benchmark calculations are provided for the new data set, which contains significantly larger molecular interactions than current databases in the literature in addition to charged fragments. Compared to density funtional theory (DFT)- and wavefunction-based methods for calculating the three-body interaction energy, our model makes predictions in a significantly shorter time frame by reducing the number of required SCF calculations from 7 to 4 performed at the PBE0 level of theory, showcasing the utility and efficiency of our Δ-ML method particularly when applied to larger systems.

Inclusion of More Physics Leads to Less Data: Learning the Interaction Energy as a Function of Electron Deformation Density with Limited Training Data.

Machine learning (ML) approaches to predicting quantum mechanical (QM) properties have made great strides toward achieving the computational chemist's holy grail of structure-based property prediction. In contrast to direct ML methods, which encode a molecule with only structural information, in this work, we show that QM descriptors improve ML predictions of dimer interaction energy, both in terms of accuracy and data efficiency, by incorporating electronic information into the descriptor. We present the electron deformation density interaction energy machine learning (EDDIE-ML) model, which predicts the interaction energy as a function of Hartree-Fock electron deformation density. We compare its performance with leading direct ML schemes and modern DFT methods for the prediction of interaction energies for dimers of varying charge type, size, and intermolecular separation. Under a low-data regime, EDDIE-ML outperforms other direct ML schemes and is the only model readily transferrable to larger, more complex systems including base pair trimers and porous cages. The underlying physical connection between the density and interaction energy enables EDDIE-ML to reach an accuracy comparable to modern DFT functionals in fewer training data points compared to other ML methods.

An ab initio Study of the Structure and Energetics of Hydrogen Bonding in Ionic Liquids.

Unlike typical hydrogen-bonded networks such as water, hydrogen bonded ionic liquids display some unusual characteristics due to the complex interplay of electrostatics, polarization, and dispersion forces in the bulk. Protic ionic liquids in particular contain close-to traditional linear hydrogen bonds that define their physicochemical properties. This work investigates whether hydrogen bonded ionic liquids (HBILs) can be differentiated from aprotic ionic liquids with no linear hydrogen bonds using state-of-the-art ab initio calculations. This is achieved through geometry optimizations of a series of single ion pairs of HBILs in the gas phase and an implicit solvent. Using benchmark CCSD(T)/CBS calculations, the electrostatic and dispersion components of the interaction energy of these systems are compared with those of aprotic ionic liquids. The inclusion of the implicit solvent significantly influenced geometries of single ion pairs, with the gas phase shortening the hydrogen bond to reduce electrostatic interactions. HBILs were found to have stronger interactions by at least 10EtMeNH0 kJ mol-1 over aprotic ILs, clearly highlighting the electrostatic nature of hydrogen bonding. Geometric and energetic parameters were found to complement each other in determining the extent of hydrogen bonding present in these ionic liquids.

Is it possible to control kinetic rates of radical polymerisation in ionic liquids?

Experimental studies have noted the often surprising and unpredictable effect of ionic liquids as solvents on reaction kinetics for radical polymerisation. We theoretically investigate the energetic and structural effects of ionic liquids, both protic and aprotic, on radical stability, presenting stabilisation of the radical by the ionic liquid by up to -78.0 kJ mol-1. Kinetic data relating to propagating systems for several industrially viable monomers indicate that propagation rates can be increased or decreased (by up to 6 orders of magnitude) depending on the monomer and ionic liquid combination. The interplay of activation entropy and activation enthalpy, much of which depends on hydrogen bonding between the solvent and reactants, play a crucial role in controlling reaction kinetics. It is concluded that the use of cheaper protic ionic liquids as solvents may be viable for improved kinetic control over radical reactions.