Yingqi Tian | Searcharxiv

arXiv2022

Mechanistic insights into water autoionization

Water autoionization plays a critical role in determining pH and properties of various chemical and biological processes occurring in the water mediated environment. The strikingly unsymmetrical potential energy surface of the dissociation process poses a great challenge to the mechanistic study. Here, we demonstrate that reliable sampling of the ionization path is accessible through nanosecond timescale metadynamics simulation enhanced by machine learning of the neural network potentials with ab initio precision, which is proved by quantitatively reproduced water equilibrium constant (p$K_\mathrm{w}$=14.14) and ionization rate constant (1.566$\times10^{-3}$ s$^{-1}$). Statistical analysis unveils the asynchronous character of the concerted triple proton transfer process. Based on conditional ensemble average calculations, we propose a dual-presolvation mechanism, which suggests that a pair of hypercoordinated and undercoordinated waters bridged by one \ce{H2O} cooperatively constitutes the initiation environment for autoionization, and contributes majorly to the local electric field fluctuation to promote water dissociation.

physics.chem-ph↗

arXiv2020

A Forecasting System of Computational Time of DFT/TDDFT Calculations under the Multiverse ansatz via Machine Learning and Cheminformatics

A top-level designed forecasting system for predicting computational times of density-functional theory (DFT)/time-dependent density-functional theory (TDDFT) calculations is presented. The computational time is assumed as the intrinsic property for the molecule. Basing on this assumption, the forecasting system is established using the "reinforced concrete", which combines the cheminformatics, several machine-learning (ML) models, and the framework of many-world interpretation (MWI) in multiverse ansatz. Herein, the cheminformatics is used to recognize the topological structure of molecules, the ML/AI models are used to build the relationships between topology and computational cost, and the MWI framework is used to hold various combinations of DFT functionals and basis sets in DFT/TDDFT calculations. Calculated results of molecules from DrugBank dataset show that 1) it can give quantitative predictions of computational costs, typical mean relative errors can be less than 0.2 for DFT/TDDFT calculations with derivations of 25% using the exactly pre-trained ML models, 2) it can also be employed to various combinations of DFT functional and basis set cases without exactly pre-trained ML models, while only slightly enlarge predicting errors.

physics.comp-ph↗