Search for: All records

Machine learning potentials (MLPs) are poised to combine the accuracy of ab initio predictions with the computational efficiency of classical molecular dynamics (MD) simulation. While great progress has been made over the last two decades in developing MLPs, there is still much to be done to evaluate their model transferability and facilitate their development. In this work, we construct two deep potential (DP) models for liquid water near graphene surfaces, Model S and Model F, with the latter having more training data. A concurrent learning algorithm (DP-GEN) is adopted to explore the configurational space beyond the scope of conventional ab initio MD simulation. By examining the performance of Model S, we find that an accurate prediction of atomic force does not imply an accurate prediction of system energy. The deviation from the relative atomic force alone is insufficient to assess the accuracy of the DP models. Based on the performance of Model F, we propose that the relative magnitude of the model deviation and the corresponding root-mean-square error of the original test dataset, including energy and atomic force, can serve as an indicator for evaluating the accuracy of the model prediction for a given structure, which is particularly applicable for large systems where density functional theory calculations are infeasible. In addition to the prediction accuracy of the model described above, we also briefly discuss simulation stability and its relationship to the former. Both are important aspects in assessing the transferability of the MLP model. 

Molecular design of redox-active materials with higher solubility and greater redox potential windows is instrumental in enhancing the performance of redox flow batteries Here we propose a computational procedure for a systematic evaluation of organic redox-active species by combining machine learning, quantum-mechanical, and classical density functional theory calculations. 1,517 small quinone molecules were generated from the building blocks of benzoquinone, naphthoquinone, and anthraquinone with different substituent groups. The physics-based methods were used to predict HOMO-LUMO gaps and solvation free energies that account for the redox potential differences and aqueous solubility, respectively. The high-throughput calculations were augmented with the quantitative structure-property relationship analyses and machine learning/graph network modeling to evaluate the materials’ overall behavior. The computational procedure was able to reproduce high-performance cathode electrolyte materials consistent with experimental observations and identify new electrolytes for RFBs by screening 100,000 di-substituted quinone molecules, the largest library of redox-active quinone molecules ever investigated. The efficient computational platform may facilitate a better understanding of the structure-function relationship of quinone molecules and advance the design and application of all-organic active materials for RFBs.