雷亚克夫
分子动力学
纳米孔
氢键
离解(化学)
密度泛函理论
化学
从头算
化学物理
计算化学
活化能
力场(虚构)
材料科学
物理化学
分子
有机化学
原子间势
人工智能
计算机科学
作者
Jessica Rimsza,Jejoon Yeon,Adri C. T. van Duin,Jincheng Du
标识
DOI:10.1021/acs.jpcc.6b07939
摘要
Detailed understanding of the reactions and processes which govern silicate–water interactions is critical to geological, materials, and environmental sciences. Interactions between water and nanoporous silica were studied using classical molecular dynamics with a Reactive Force Field (ReaxFF), and the results were compared with density functional theory (DFT) based ab initio molecular dynamics (AIMD) simulations. Two versions of ReaxFF Si/O/H parametrizations (Yeon et al. J. Phys. Chem. C 2016, 120, 305 and Fogarty et al. J. Chem. Phys. 2010, 132, 174704) were compared with AIMD results to identify differences in local structures, water dissociation mechanisms, energy barriers, and diffusion behaviors. Results identified reaction mechanisms consisting of two different intermediate structures involved in the removal of high energy two-membered ring (2-Ring) defects on complex nanoporous silica surfaces. Intermediate defects lifetimes affect hydroxylation and 2-Ring defect removal. Additionally, the limited internal volume of the nanoporous silica results in decreased water diffusion related to the development of nanoconfined water. Hydrogen atoms in the water diffused 10–30% faster than the oxygen atoms, suggesting that increased hydrogen diffusion through hydrogen hopping mechanisms may be enhanced in nanoconfined conditions. Comparison of the two different ReaxFF parametrizations with AIMD data indicated that the Yeon et al. parameters resulted in reaction mechanisms, hydroxylation rates, defect concentrations, and activation energies more consistent with the AIMD simulations. Therefore, this ReaxFF parametrization is recommended for future studies of water–silica systems with high concentrations of surface defects and highly strained siloxane bonds such as in complex silica nanostructures.
科研通智能强力驱动
Strongly Powered by AbleSci AI