Poly(ethylene oxide) (PEO)-based polymer electrolytes are a promising class of materials for use in lithium-ion batteries due to their high ionic conductivity and flexibility. In this study, the effects of polymer architecture including linear, star, and hyperbranched and salt (lithiumbis(trifluoromethanesulfonyl)imide (LiTFSI)) concentration on the glass transition (Tg), microstructure, phase diagram, free volume, and bulk viscosity, all of which play a significant role in determining the ionic conductivity of the electrolyte, have been systematically studied for PEO-based polymer electrolytes. The branching of PEO widens the liquid phase toward lower salt concentrations, suggesting decreased crystallization and improved ion coordination. At high salt loadings, ion clustering is common for all electrolytes, yet the cluster size and distribution appear to be strongly architecture-dependent. Also, the ionic conductivity is maximized at a salt concentration of [Li/EO ≈ 0.085] for all architectures, and the highly branched polymers displayed as much as three times higher ionic conductivity (with respect to the linear analogue) for the same total molar mass. The architecture-dependent ionic conductivity is attributed to the enhanced free volume measured by positron annihilation lifetime spectroscopy. Interestingly, despite the strong architecture dependence of ionic conductivity, the salt addition in the highly branched architectures results in accelerated yet similar monomeric friction coefficients for these polymers, offering significant potential toward decoupling of conductivity from segmental dynamics of polymer electrolytes, leading to outstanding battery performance.
The Major Histocompatibility Complex (MHC) binds to the derived peptides from pathogens to present them to killer T cells on the cell surface. Developing computational methods for accurate, fast, and explainable peptide-MHC binding prediction can facilitate immunotherapies and vaccine development. Various deep learning-based methods rely on separate feature extraction from the peptide and MHC sequences and ignore their pairwise binding information. This paper develops a capsule neural network-based method to efficiently capture the peptide-MHC complex features to predict the peptide-MHC class I binding. Various evaluations confirmed our method outperformance over the alternative methods, while it can provide accurate prediction over less available data. Moreover, for providing precise insights into the results, we explored the essential features that contributed to the prediction. Since the simulation results demonstrated consistency with the experimental studies, we concluded that our method can be utilized for the accurate, rapid, and interpretable peptide-MHC binding prediction to assist biological therapies.