基于Zhu-Nakamura跃迁几率算法的非绝热直接动力学模拟方法

Internal conversion and intersystem crossing are ubiquitous nonadiabatic processes in photochemistry and photobiology, and which are of fundamental importance. The classical trajectory based surface hopping methodology has become one of the main tools for the nonadiabatic processes, and the key procedure in which is the calculation of nonadiabatic transition probability. This project will present a new scheme based on the Zhu-Nakamura theory that could be used to estimating the intersystem crossing probability for intersected or quasi degenerated states with different spin multiplicities. The new scheme employs the two-step or three-step model in locating the crossing region. The intersystem crossing probability calculation and the velocity adjustment after hop are evaluated by only two adiabatic potential energy surfaces and its gradients and spin-orbit coupling matrix elements. The new scheme and the existed internal conversion probability calculation method are based on the unified framework, and enable them generally applicable to noadiabatic transition problems between states of the same or different spin multiplicities in photochemical reaction processes. Based on this scheme, the project will develop new nonadiabatic on-the-fly dynamics programs, which will be able to deal with real time dynamic simulation for large systems or complicated photochemical processes involve multi-electronic-states.

内转换和系间窜越是在光化学和光生物反应中广泛存在的非绝热过程，具有重要的基础意义。基于经典轨线的面跳跃动力学成为研究非绝热过程的主要方法之一，其核心内容是非绝热跃迁几率的计算。本项目基于Zhu-Nakamura理论提出一个计算不同自旋电子态间交叉和准简并情形下系间窜越几率的新算法。新算法使用两点和三点模型寻找交叉点，并且只需要交叉区域势能、能量梯度合旋轨耦合矩阵元即可计算系间窜越几率并处理跃迁后分子运动调整。新算法与已有的内转换几率算法理论框架一致，两者结合可以通用化地处理光化学反应过程中相同和不同自旋电子态间非绝热跃迁。基于此算法，本项目拟发展一种新的非绝热直接动力学程序，为大分子体系或涉及多个电子态的复杂光化学反应动力学实时模拟提供有力的计算支持。

本项目的研究目标是发展基于Zhu-Nakamura公式的系间窜越几率算法并将其与现有的内转换几率算法结合，构建可同步处理圆锥交叉、避免交叉型系间窜越和准简并型系间窜越非绝热跃迁几率的算法体系和从头算动力学模拟程序开发。截止项目结题，上述目标均已实现并在示例光反应体系上进行了测试。此外,我们还在以下方面对模拟程序进行了改进：制定标准化输入和输出文件、随轨线演化实时计算SOC、简并型避免交叉区域的跃迁点搜索、同位素效应模拟、适用于研究多步反应的产物分离方法。本项目首次提出的改进的Zhu-Nakamura跃迁几率算法仅需随轨线演化获取势能、能量梯度和旋轨耦合矩阵元即可计算分子在当前构型由所处电子态向相邻自旋相同或不同电子态的跃迁几率，克服了传统的fewest-switch方法对非绝热耦合矩阵元的依赖,并且在精度水平一致的前提下我们的方法具有更加优越的计算效率。本项目所发展的方法和程序为开展更大分子体系、高激发电子态，不同自旋电子态和多通道竞争反应过程的动力学理论研究铺平了道路。