化学反应速率常数的确定，在燃烧、大气以及星际化学反应的化学动力学建模中起着不可或缺的作用，为建模提供了重要参数。而实验和常用的理论如量子散射理论和过渡态理论均难以准确并快速地确定多原子反应的速率常数。本项目拟采用近年来新发展的全维量子动力学理论，ring-polymer分子动力学（RPMD），并借助经典分子动力学中的先进控温技术和抽样算法，应用基于并行计算和GPU加速技术优化后的高效率计算程序，来模拟一系列在过渡态两侧均存在显著范德华势阱的自由基加水的化学反应，X+H2O->XH+HO（X=O, F, Cl, Br），系统地大规模计算该反应在不同温度下与各种同位素的反应速率常数和动力学同位素效应。本工作将在验证RPMD在化学反应速率常数计算中的有效性的同时，也进一步考察范德华势阱、隧穿效应与零点能等量子效应，以及非绝热效应对此类反应的影响，从而揭示此类反应的动力学特点。

The determination of chemical reaction coefficients plays central role in the chemical kinetics modeling of combustion, atmospheric and interstellar chemistry by offering key parameters. But it’s still formidable to confirm the rate coefficients for poly-atom reactions accurately and efficiently, either via experiment or popular theories such as quantum scattering theory or transition state theory. In this project, we try to exploit a recently developed full dimensional quantum dynamics, ring-polymer molecular dynamics (RPMD), with advanced algorithms from classical molecular dynamics in both thermostat and enhanced sampling, and optimized by both parallel computing and GPU technology, to extensively calculate the rate coefficients and kinetic isotope effects at different temperatures for a series of reactions, X+H2O->XH+HO (X=O, F, Cl, Br), which are featured by having the van der Waals wells at both sides of the transition states. We plan to not only validate the accuracy and efficiency of RPMD in rate coefficient calculations, but also investigate the influence of van der Waals wells, quantum effects such as tunneling and zero point energy, and non-adiabatic effects for the title reaction. The ultimate goal is to reveal the dynamic characteristics for the title reactions.