Machine Learning Solvation Environments in Conductive Polymers: Application to ProDOT-2Hex with Solvent Swelling

Abstract

<div>Automated identification and classification of ion solvation sites in diverse chemical systems will improve the understanding and design of polymer electrolytes for battery applications. We introduce a machine learning approach to classify and characterize ion solvation environments based on feature vectors extracted from all-atom simulations. This approach is demonstrated in poly(3,4-propylenedioxythiophene), which is a promising candidate polymer binder for Li-ion batteries. In the dry polymer, four</div><div>distinct Li+ solvation environments are identified close to the backbone of the polymer. Upon swelling of the polymer with propylene carbonate solvent, the nature of Li+ solvation changes dramatically, featuring a rapid diversification</div><div>of solvation environments. This application of machine learning can be generalized to other polymer condensed-phase systems to elucidate the molecular mechanisms underlying ion solvation.</div>

A review of the recent progress in battery informatics

“…Conventionally, the analysis of molecular dynamics trajectory is carried out through hand-crafted rules in combination with computing the average behavior of atoms. The powerful capability of ML in handling a large amount of data opens new opportunities to post-analyze the MD data to mitigate potential information loss during the analysis 137,[195][196][197][198] . Particularly, an interesting application of ML in analyzing MD data is labeling atoms in distinct coordination environments through unsupervised clustering.…”

“…The difference of site availability reflects the conduction characteristics in these two phases. Magdau and Miller developed a machine leaning approach to automate the classification and identification of ion solvation environments in polymer electrolyte based on data from MD simulations 197 . By concatenating the typespecific Li + radial distribution functions, they applied two unsupervised algorithms of UMAP to embed the high dimensional feature vectors into a low-dimensional latent space and HDBSCAN to classify the embedded data into specific solvating environments in poly (3,4-propylenedioxythiophene).…”

Machine learning based implicit solvent model for aqueous-solution alanine dipeptide molecular dynamics simulations

“…Their pioneering work follows the recent, rapidly growing use of ML techniques 51 in the development of computational chemistry models. [52][53][54] As far as the solvent effect is concerned, ML models have been developed to capture the effect on chemical reactions, 55,56 spectral properties, [57][58][59][60] identify solvation characteristics in general molecular environment, 50,61,62 and explore the solvent effect on mixture solvent system. [63][64][65] Inspired by the work of Noé and coworkers, 42,49,50,66 in this paper we will also use the solvated alanine dipeptide [67][68][69] as an example and further explore the possibility of "deriving" an implicit solvent model directly from explicit solvent MD simulations.…”

Accurate Molecular-Orbital-Based Machine Learning Energies via Unsupervised Clustering of Chemical Space

“…Machine learning (ML) approaches have attracted considerable interest in the chemical sciences for a variety of applications, including molecular and material design, [1][2][3][4][4][5][6][7][8] protein property prediction, [9][10][11] reaction mechanism discovery, 4,[12][13][14][15][16] and analysis and classification tasks for new physical insights. [17][18][19] As an alternative to physics-based computations, ML has also shown promise for the prediction of molecular energies, [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] intermolecular interactions, 31,35 electron densities, 21,24,[36][37][38] and linear response properties. [39][40][41][42]…”