复杂分子体系的相空间量子动力学理论及高效算法

Quantum time correlation functions, which bridge macroscopic physical observables (spectra, reaction rates, etc.) on dynamics and microscopic dynamic pictures on the atomic or molecular level, present the most important and powerful tool to understand, describe, and predict quantum effects in dynamic processes of complex molecular systems. The theoretical method for faithfully studying chemical dynamics of thermal equilibrium complex systems should be able to conserve the quantum Boltzmann distribution during propagation and to approach exact results of any kinds of quantum time correlation functions in the classical, high-temperature, and harmonic limits. While most practical theoretical methods or models for complex systems are not able to simultaneously meet the two requirements based on quantum dynamics and quantum statistics, we have recently independently developed a novel theoretical framework on dynamics in phase space formulations of quantum mechanics which can fulfil both of them. The present proposal aims to further develop adiabatic approximations for equations of motion and enhanced sampling methods in our proposed theoretical framework, which lead to practical approaches for complex molecular systems and their efficient parallel computing programs. The results will enhance the accuracy and prediction power of chemical dynamics methods for experimental observables and for dynamic behaviors of complex molecular system.

量子时间关联函数将宏观动力学物理量（光谱、反应速率等）和原子分子层面的微观动态绘景直接联系起来，是理解、描述和预测复杂分子体系动力学过程中量子效应的最重要、最有效的手段。准确描述热平衡复杂体系化学动力学的理论方法应当保证量子玻尔兹曼分布函数随时间的不变性，而且对任意量子时间关联函数都能给出准确的经典极限、高温极限和谐振子极限。已有的适用于复杂体系的理论方法不能同时满足这两个基于动力学方程和量子统计力学的重要要求。申请人最近独立发展的相空间量子动力学的一套新的理论框架突破了原有理论方法的局限。本项目拟在此基础上探索量子相空间表达形式的优化选项，进一步发展适用于复杂分子体系的绝热近似方法和增强采样方法，建立一套完整的可并行化的高效计算程序，大大提升化学动力学方法对复杂体系动态行为和实验观测量的计算精度和预言能力。

量子时间关联函数将宏观化学动力学物理量（光谱、反应速率等）和原子分子层面的微观绘景直接联系起来，是理解、描述和预测复杂分子体系动力学过程中量子效应的最重要、最有效的手段。用来描述热平衡复杂体系化学动力学的理论方法应当保证量子玻尔兹曼分布函数随时间的不变性，而且对任意量子时间关联函数都能给出准确的经典极限、高温极限和谐振子极限。本项目取得的进展如下.1）独立发展的路径积分刘维尔动力学方法突破近二十年来理论方法的局限，能同时满足这两个基于动力学方程和量子统计力学的重要要求。.2）独立发展的针对路径积分分子动力学的高效热库采样框架，将效率提高了5~10倍。.本项目同时建立了一套完整的可并行化的高效计算程序，提升了对于复杂分子体系化学动力学和量子统计力学的计算精度和预言能力。