AI 中文总结
该研究开发的非绝热前向通量采样（NAFFS）方法可高效模拟全维分子系统中的稀有非绝热过程，以气相氨光解离为基准验证，其采样的反应轨迹统计权重正确，模拟时间大幅缩短。
AI 中文摘要
通过蛮力表面跳变模拟光化学中的稀有事件在计算上是不可行的：大多数传播轨迹要么保持非反应性，要么遵循主导的弛豫通道，而目标反应的发生概率极低。路径采样方法长期以来用于解决基态动力学中的这一问题，但将其扩展到非绝热过程迄今仅局限于低维分析模型。本文证明，我们新近开发的非绝热前向通量采样（NAFFS）方法可用于高效模拟全维分子系统中的稀有事件。作为代表性基准，我们研究气相氨的稀有分子光解离通道：NH₃ + hν → NH + H₂。NAFFS以正确的统计权重采样反应轨迹，重现了此前蛮力动力学研究得到的反应速率常数，可通过平均承诺子分析直接获取机理信息，并将每条反应轨迹的模拟时间减少了三个数量级。这些结果确立了NAFFS作为高效且统计严谨的框架，用于研究现实分子系统中的稀有非绝热过程。
英文摘要
Simulating rare events in photochemistry by brute-force surface hopping is computationally prohibitive: most propagated trajectories either remain nonreactive or follow dominant relaxation channels, while the reaction of interest may occur with very low probability. Path sampling methods have long addressed this problem for ground-state dynamics, but their extension to nonadiabatic processes has so far been limited to low-dimensional analytical models. Here, we demonstrate that our recently developed nonadiabatic forward-flux sampling (NAFFS) method can be applied to efficiently simulate rare-events in full-dimensional molecular systems. As a representative benchmark, we investigate the rare molecular photodissociation channel of gas-phase ammonia, NH$_3$ + h$ν\rightarrow$ NH + H$_2$. NAFFS samples reactive trajectories with the correct statistical weight, reproduces the reaction rate constant obtained in previous brute-force dynamical studies, provides direct access to mechanistic information through an averaged committor analysis, and reduces the simulation time per transition trajectory by up to three orders of magnitude. These results establish NAFFS as an efficient and statistically rigorous framework for investigating rare nonadiabatic processes in realistic molecular systems.
