Temperature-Dependent Density and Viscosity Prediction for Hydrocarbons: Machine Learning and Molecular Dynamics Simulations.

Machine learning-based predictive models allow rapid and reliable prediction of material properties and facilitate innovative materials design. Base oils used in the formulation of lubricant products are complex hydrocarbons of varying sizes and structure. This study developed Gaussian process regression-based models to accurately predict the temperature-dependent density and dynamic viscosity of 305 complex hydrocarbons. In our approach, strongly correlated/collinear predictors were trimmed, important predictors were selected by least absolute shrinkage and selection operator (LASSO) regularization and prior domain knowledge, hyperparameters were systematically optimized by Bayesian optimization, and the models were interpreted. The approach provided versatile and quantitative structure-property relationship (QSPR) models with relatively simple predictors for determining the dynamic viscosity and density of complex hydrocarbons at any temperature. In addition, we developed molecular dynamics simulation-based descriptors and evaluated the feasibility and versatility of dynamic descriptors from simulations for predicting the material properties. It was found that the models developed using a comparably smaller pool of dynamic descriptors performed similarly in predicting density and viscosity to models based on many more static descriptors. The best models were shown to predict density and dynamic viscosity with coefficient of determination (R2) values of 99.6% and 97.7%, respectively, for all data sets, including a test data set of 45 molecules. Finally, partial dependency plots (PDPs), individual conditional expectation (ICE) plots, local interpretable model-agnostic explanation (LIME) values, and trimmed model R2 values were used to identify the most important static and dynamic predictors of the density and viscosity.

Simulations of Friction Anisotropy on Self-Assembled Monolayers in Water.

Molecular dynamics simulations were performed to study nanoscale friction on hydrophilic and hydrophobic self-assembled monolayers (SAMs) immersed in water. Sliding was simulated in two different directions to capture anisotropy due to the direction of motion relative to the inherent tilted orientation of the molecules. It was shown that friction depends on both hydrophobicity and sliding direction, with the highest friction observed for sliding on hydrophobic SAM in the direction against the initial orientation of the molecules. The origins of the friction trends were analyzed by differentiating the tip-SAM and tip-water force contributions to friction. The tip-water force was higher on the hydrophilic SAM, and this was shown to be due to the presence of a dense layer of water adjacent to the surface and hydrogen bonding. In contrast, the tip-SAM force was higher on the hydrophobic SAM due to a water depletion layer, which enabled the tip to be closer to the SAM terminal group. The higher-friction cases all exhibited greater penetration of the tip below the surface of the SAM, accommodated by further tilting and reorientation of the SAM molecules.

Simulation of Subnanometer Contrast in Dynamic Atomic Force Microscopy of Hydrophilic Alkanethiol Self-Assembled Monolayers in Water.

Atomic resolution imaging of surfaces in liquid environments using atomic force microscopy (AFM) is challenging in terms of both reproducibility and measurement interpretation. To understand the origins of these challenges, we used molecular dynamics simulations of AFM on hydrophilic self-assembled monolayers (SAMs) in water. The force on the model AFM tip was calculated as a function of lateral and vertical position relative to the SAM surface. The contributions of the water and SAMs to the overall force were analyzed, and the former was correlated to the water density distribution. Then, dynamic AFM was modeled by oscillating the tip at a driving amplitude. It was found that the contrast between amplitudes at different lateral positions on the surface was dependent on the vertical position of the tip. Lastly, amplitude maps were produced for two vertical positions at constant height, and the ability to capture atomic resolution was related to the force on the tip. These results offer an explanation for the observed instability in atomic scale imaging using AFM and more generally provide insight into the contrast mechanisms of surface images obtained in liquid environments.

Self-assembly of nanocrystal checkerboard patterns via non-specific interactions.

Checkerboard lattices-where the resulting structure is open, porous, and highly symmetric-are difficult to create by self-assembly. Synthetic systems that adopt such structures typically rely on shape complementarity and site-specific chemical interactions that are only available to biomolecular systems (e.g., protein, DNA). Here we show the assembly of checkerboard lattices from colloidal nanocrystals that harness the effects of multiple, coupled physical forces at disparate length scales (interfacial, interparticle, and intermolecular) and that do not rely on chemical binding. Colloidal Ag nanocubes were bi-functionalized with mixtures of hydrophilic and hydrophobic surface ligands and subsequently assembled at an air-water interface. Using feedback between molecular dynamics simulations and interfacial assembly experiments, we achieve a periodic checkerboard mesostructure that represents a tiny fraction of the phase space associated with the polymer-grafted nanocrystals used in these experiments. In a broader context, this work expands our knowledge of non-specific nanocrystal interactions and presents a computation-guided strategy for designing self-assembling materials.

PyL3dMD: Python LAMMPS 3D molecular descriptors package.

Molecular descriptors characterize the biological, physical, and chemical properties of molecules and have long been used for understanding molecular interactions and facilitating materials design. Some of the most robust descriptors are derived from geometrical representations of molecules, called 3-dimensional (3D) descriptors. When calculated from molecular dynamics (MD) simulation trajectories, 3D descriptors can also capture the effects of operating conditions such as temperature or pressure. However, extracting 3D descriptors from MD trajectories is non-trivial, which hinders their wide use by researchers developing advanced quantitative-structure-property-relationship models using machine learning. Here, we describe a suite of open-source Python-based post-processing routines, called PyL3dMD, for calculating 3D descriptors from MD simulations. PyL3dMD is compatible with the popular simulation package LAMMPS and enables users to compute more than 2000 3D molecular descriptors from atomic trajectories generated by MD simulations. PyL3dMD is freely available via GitHub and can be easily installed and used as a highly flexible Python package on all major platforms (Windows, Linux, and macOS). A performance benchmark study used descriptors calculated by PyL3dMD to develop a neural network and the results showed that PyL3dMD is fast and efficient in calculating descriptors for large and complex molecular systems with long simulation durations. PyL3dMD facilitates the calculation of 3D molecular descriptors using MD simulations, making it a valuable tool for cheminformatics studies.

Probabilistic Approach to Low Strain Rate Atomistic Simulations of Ultimate Tensile Strength of Polymer Crystals.

Molecular dynamics simulations of the tensile ultimate properties of polymer crystals require the use of empirical potentials that model bond dissociation. However, fully reactive potentials are computationally expensive such that reactive simulations cannot reach the low strain rates of typical experiments. Here, we present a hybrid approach that uses the simplicity of a classical, nonreactive potential, information from bond dissociation energy calculations, and a probabilistic expression that mimics bond breaking. The approach is demonstrated for poly(p-phenylene terephthalamide) and, with one tunable parameter, the calculated tensile ultimate stress matches that obtained using a fully reactive simulation at high strain rates. Then, the hybrid simulations are run at much lower strain rates where the ultimate tensile stress is strain rate-independent and consistent with the expected experimental range.

Methanol-Assisted ADMET Polymerization of Semiaromatic Amides.

A method for the acyclic diene metathesis polymerization of semiaromatic amides is described. The procedure uses second-generation Grubbs' catalyst and N-cyclohexyl-2-pyrrolidone (CHP), a high boiling, polar solvent capable of solubilizing both monomer and polymer. The addition of methanol to the reaction was found to significantly increase polymer molar mass although the role of the alcohol is currently not understood. Hydrogenation with hydrogen gas and Wilkinson's catalyst resulted in near-quantitative saturation. All polymers synthesized here exhibit a hierarchical semicrystalline morphology driven by ordering of aromatic amide groups via strong nonbonded interactions. Furthermore, the melting points can be tuned over a >100 °C range by precise substitution at just one of the backbone positions on each mer (<5% of the total).