离子水溶液的氢键网络结构和动力学研究

溶液中的水合离子与生物大分子在系列生化反应中的行为密切相关，并遵循一定的离子特异性，如Hofmeister序列。但目前人们对于Hofmeister序列存在的原因尚不清楚。相关的衍射散射光谱实验、激发态光谱实验等对于溶液的研究存在较多争议，主要集中在离子对水的氢键网络结构和动力学的影响能否超出最近邻的水合层，以及溶液中结晶成核过程的微观机制等。而理论上，经验力场难以处理溶液中复杂极化环境、量子效应和氢键作用，第一性原理计算模拟大尺度溶液体系效率低。针对上述问题和困难，本项目将利用基于机器学习对溶液进行高精度高效率的深度势能分子动力学模拟，其深度学习力场来自于训练第一性原理计算的数据，重点探讨离子对水的结构影响的作用范围，以及对水的动力学性质的影响。该项目将为解释实验现象和阐明争议问题提供理论基础，并为进一步分析多种类阴阳离子之间的协同效应、设计海水净化功能材料等提供有意义的参考。

In solution, hydrated ion plays a critical role in the behavior of biomolecules in various biochemical processes, which follows ion-specific series, such as the Hofmeister series. However, our knowledge of the Hofmeister series at the molecular level remain elusive, especially the anomalies and cation-anion cooperativity, etc. There are still some controversies in recent scattering and diffraction spectroscopy experiments of aqueous solutions, mainly in the extent to which structural changes extend outside the first solvation shell, as well as the mechanism of nucleation and crystallization processes. Theoretically, classical force field hardly deals with complex polarized environment, quantum effects and hydrogen bonding. Meanwhile, the computational cost of ab initio calculation limits its application to large aqueous system. To address the above issues, we will employ the machine learning method to perform accurate and efficient deep-potential molecular dynamics simulations of solution, where the deep learning force field is obtained by training results from ab initio calculation. We will focus on the understanding the local and global effect of ions on the structure and dynamics of water. This project will explain experimental observations, clarify previous controversies, and provide helpful results for further understanding of cation-anion cooperativity and water desalination material design.